Rui Qian, Chuanhang Deng, Qiang Huang, Jian Xiong, Mingxuan Li, Yingbo Zhou, Wei Zhai, Jintao Chen, Dejing Dou. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Reinforcement learning (RL) has emerged as an effective approach for enhancing the reasoning capabilities of large language models (LLMs), especially in scenarios where supervised fine-tuning (SFT) falls short due to limited chain-of-thought (CoT) data.Among RLbased post-training methods, group relative advantage estimation, as exemplified by Group Relative Policy Optimization (GRPO), has attracted considerable attention for eliminating the dependency on the value model, thereby simplifying training compared to traditional approaches like Proximal Policy Optimization (PPO).However, existing group relative advantage estimation method still suffers from training inefficiencies, particularly when the estimated advantage approaches zero.To address this limitation, we propose Advantage-Augmented Policy Optimization (AAPO), a novel RL algorithm that optimizes the crossentropy (CE) loss using advantages enhanced through a margin-based estimation scheme.This approach effectively mitigates the inefficiencies associated with group relative advantage estimation.Experimental results on multiple mathematical reasoning benchmarks and model series demonstrate the superior performance of AAPO.

Current Large Multimodal Models (LMMs) empowered visual grounding typically rely ontokens as a text prompt to jointly optimize the vision-language model (e.g., LLaVA) and the downstream task-specific model (e.g., SAM). However, we observe that little research has looked into how it works. In this work, we first visualize the similarity maps, which are obtained by computing the semantic similarity between thetoken and the image token embeddings derived from the last hidden layer in both the LLaVA encoder and SAM decoder. Intriguingly, we have found that a striking consistency holds in terms of activation responses in the similarity map, which reveals that what thetoken contributes to is semantic similarity within image-text pairs. Specifically, thetoken, a placeholder expanded in text vocabulary, extensively queries among individual tokenized image patches to match the semantics of an object from text to the paired image, while the Large Language Models (LLMs) are being fine-tuned. Upon the above findings, we present READ, which facilitates LMMs’ resilient REAsoning capability of where to attenD under the guidance of highly activated points borrowed from similarity maps. Remarkably, READ features an intuitive design, Similarity as Points module (SasP), which can be seamlessly applied to-like paradigms in a plug-and-play fashion. Also, extensive experiments have been conducted on ReasonSeg and RefCOCO(+/g) datasets. To validate whether READ suffers from catastrophic forgetting of previous skills after fine-tuning, we further assess its generation ability on an augmented FP-RefCOCO(+/g) dataset. All codes and models are publicly available at https://github.com/rui-qian/READ.

Machine unlearning (MU) aims to remove the influence of specific data points from trained models, enhancing compliance with privacy regulations. However, the vulnerability of basic MU models to malicious unlearning requests in adversarial learning environments has been largely overlooked. Existing adversarial MU attacks suffer from three key limitations: inflexibility due to pre-defined attack targets, inefficiency in handling multiple attack requests, and instability caused by non-convex loss functions. To address these challenges, we propose a Flexible, Efficient, and Stable Attack (DDPA). First, leveraging Carathéodory's theorem, we introduce a convex polyhedral approximation to identify points in the loss landscape where convexity approximately holds, ensuring stable attack performance. Second, inspired by simplex theory and John's theorem, we develop a regular simplex detection technique that maximizes coverage over the parameter space, improving attack flexibility and efficiency. We theoretically derive the proportion of the effective parameter space occupied by the constructed simplex. We evaluate the attack success rate of our DDPA method on real datasets against state-of-the-art machine unlearning attack methods. Our source code is available at https://github.com/zzz0134/DDPA.

The restoration of hyperspectral image (HSI) plays a pivotal role in subsequent hyperspectral image applications. Despite the remarkable capabilities of deep learning, current HSI restoration methods face challenges in effectively exploring the spatial non-local self-similarity and spectral low-rank property inherently embedded with HSIs. This paper addresses these challenges by introducing a latent diffusion enhanced rectangle Transformer for HSI restoration, tackling the non-local spatial similarity and HSI-specific latent diffusion low-rank property. In order to effectively capture non-local spatial similarity, we propose the multi-shape spatial rectangle self-attention module in both horizontal and vertical directions, enabling the model to utilize informative spatial regions for HSI restoration. Meanwhile, we propose a spectral latent diffusion enhancement module that generates the image-specific latent dictionary based on the content of HSI for low-rank vector extraction and representation. This module utilizes a diffusion model to generatively obtain representations of global low-rank vectors, thereby aligning more closely with the desired HSI. A series of comprehensive experiments were carried out on four common hyperspectral image restoration tasks, including HSI denoising, HSI super-resolution, HSI reconstruction, and HSI inpainting. The results of these experiments highlight the effectiveness of our proposed method, as demonstrated by improvements in both objective metrics and subjective visual quality.

Federated Learning (FL) is a promising distributed machine learning approach that enables collaborative training of a global model using multiple edge devices. The data distributed among the edge devices are highly heterogeneous. Thus, FL faces the challenge of data distribution and heterogeneity, where non-Independent and Identically Distributed (non-IID) data across edge devices may yield in significant accuracy drop. Furthermore, the limited computation and communication capabilities of edge devices increase the likelihood of stragglers, thus leading to slow model convergence. In this article, we propose the FedDHAD FL framework, which comes with two novel methods: dynamic heterogeneous model aggregation (FedDH) and adaptive dropout (FedAD). FedDH dynamically adjusts the weights of each local model within the model aggregation process based on the non-IID degree of heterogeneous data to deal with the statistical data heterogeneity. FedAD performs neuron-adaptive operations in response to heterogeneous devices to improve accuracy while achieving superb efficiency. The combination of these two methods makes FedDHAD significantly outperform state-of-the-art solutions in terms of accuracy (up to 6.7% higher), efficiency (up to 2.02 times faster), and computation cost (up to 15.0% smaller).

The advent of large-scale pretrained language models (PLMs) has contributed greatly to the progress in natural language processing (NLP). Despite its recent success and wide adoption, fine-tuning a PLM often suffers from overfitting, which leads to poor generalizability due to the extremely high complexity of the model and the limited training samples from downstream tasks. To address this problem, we propose a novel and effective fine-tuning framework, named layerwise noise stability regularization (LNSR). Specifically, our method perturbs the input of neural networks with the standard Gaussian or in-manifold noise in the representation space and regularizes each layer's output of the language model. We provide theoretical and experimental analyses to prove the effectiveness of our method. The empirical results show that our proposed method outperforms several state-of-the-art algorithms, such as norm and start point (L2-SP), Mixout, FreeLB, and smoothness inducing adversarial regularization and Bregman proximal point optimization (SMART). In addition to evaluating the proposed method on relatively simple text classification tasks, similar to the prior works, we further evaluate the effectiveness of our method on more challenging question-answering (QA) tasks. These tasks present a higher level of difficulty, and they provide a larger amount of training examples for tuning a well-generalized model. Furthermore, the empirical results indicate that our proposed method can improve the ability of language models to domain generalization.

To explain the prediction result of a Deep Neural Network (DNN) model based on a given sample, LIME [1] and its derivatives have been proposed to approximate the local behavior of the DNN model around the data point via linear surrogates. Though these algorithms interpret the DNN by finding the key features used for classification, the random interpolations used by LIME would perturb the explanation result and cause the instability and inconsistency between repetitions of LIME computations. To tackle this issue, we propose G-LIME that extends the vanilla LIME through high-dimensional Bayesian linear regression using the sparsity and informative global priors. Specifically, with a dataset representing the population of samples (e.g., the training set), G-LIME first pursues the global explanation of the DNN model using the whole dataset. Then, with a new data point, -LIME incorporates an modified estimator of ElasticNet-alike to refine the local explanation result through balancing the distance to the global explanation and the sparsity/feature selection in the explanation. Finally, G-LIME uses Least Angle Regression (LARS) and retrieves the solution path of a modified ElasticNet under varying -regularization, to screen and rank the importance of features [2] as the explanation result. Through extensive experiments on real world tasks, we show that the proposed method yields more stable, consistent, and accurate results compared to LIME.

As a promising approach to deal with distributed data, Federated Learning (FL) achieves major advancements in recent years. FL enables collaborative model training by exploiting the raw data dispersed in multiple edge devices. However, the data is generally non-independent and identically distributed, i.e., statistical heterogeneity, and the edge devices significantly differ in terms of both computation and communication capacity, i.e., system heterogeneity. The statistical heterogeneity leads to severe accuracy degradation while the system heterogeneity significantly prolongs the training process. In order to address the heterogeneity issue, we propose an Asynchronous Staleness-aware Model Update FL framework, i.e., FedASMU, with two novel methods. First, we propose an asynchronous FL system model with a dynamical model aggregation method between updated local models and the global model on the server for superior accuracy and high efficiency. Then, we propose an adaptive local model adjustment method by aggregating the fresh global model with local models on devices to further improve the accuracy. Extensive experimentation with 6 models and 5 public datasets demonstrates that FedASMU significantly outperforms baseline approaches in terms of accuracy (0.60% to 23.90% higher) and efficiency (3.54% to 97.98% faster).

As a promising paradigm to collaboratively train models with decentralized data, Federated Learning (FL) can be exploited to fine-tune Large Language Models (LLMs).While LLMs correspond to huge size, the scale of the training data significantly increases, which leads to tremendous amounts of computation and communication costs.The training data is generally non-Independent and Identically Distributed (non-IID), which requires adaptive data processing within each device.Although Low-Rank Adaptation (LoRA) can significantly reduce the scale of parameters to update in the fine-tuning process, it still takes unaffordable time to transfer the low-rank parameters of all the layers in LLMs.In this paper, we propose a Fisher Information-based Efficient Curriculum Federated Learning framework (FibecFed) with two novel methods, i.e., adaptive federated curriculum learning and efficient sparse parameter update.First, we propose a fisher informationbased method to adaptively sample data within each device to improve the effectiveness of the FL fine-tuning process.Second, we dynamically select the proper layers for global aggregation and sparse parameters for local update with LoRA so as to improve the efficiency of the FL fine-tuning process.Extensive experimental results based on 10 datasets demonstrate that FibecFed yields excellent performance (up to 45.35% in terms of accuracy) and superb fine-tuning speed (up to 98.61% faster) compared with 17 baseline approaches).Our code will be publicly available.

Federated Learning (FL) has achieved significant achievements recently, enabling collaborative model training on distributed data over edge devices. Iterative gradient or model exchanges between devices and the centralized server in the standard FL paradigm suffer from severe efficiency bottlenecks on the server. While enabling collaborative training without a central server, existing decentralized FL approaches either focus on the synchronous mechanism that deteriorates FL convergence or ignore device staleness with an asynchronous mechanism, resulting in inferior FL accuracy. In this paper, we propose an Asynchronous Efficient Decentralized FL framework, i.e., AEDFL, in heterogeneous environments with three unique contributions. First, we propose an asynchronous FL system model with an efficient model aggregation method for improving the FL convergence. Second, we propose a dynamic staleness-aware model update approach to achieve superior accuracy. Third, we propose an adaptive sparse training method to reduce communication and computation costs without significant accuracy degradation. Extensive experimentation on four public datasets and four models demonstrates the strength of AEDFL in terms of accuracy (up to 16.3% higher), efficiency (up to 92.9% faster), and computation costs (up to 42.3% lower).

Summary While data is distributed in multiple edge devices, federated learning (FL) is attracting more and more attention to collaboratively train a machine learning model without transferring raw data. FL generally exploits a parameter server and a large number of edge devices during the whole process of the model training, while several devices are selected in each round. However, straggler devices may slow down the training process or even make the system crash during training. Meanwhile, other idle edge devices remain unused. As the bandwidth between the devices and the server is relatively low, the communication of intermediate data becomes a bottleneck. In this article, we propose time‐efficient asynchronous federated learning with sparsification and quantization, that is, TEASQ‐Fed. TEASQ‐Fed can fully exploit edge devices to asynchronously participate in the training process by actively applying for tasks. We utilize control parameters to choose an appropriate number of parallel edge devices, which simultaneously execute the training tasks. In addition, we introduce a caching mechanism and weighted averaging with respect to model staleness to further improve the accuracy. Furthermore, we propose a sparsification and quantitation approach to compress the intermediate data to accelerate the training. The experimental results reveal that TEASQ‐Fed improves the accuracy (up to 16.67% higher) while accelerating the convergence of model training (up to twice faster).

Abstract Fine-tuning deep neural networks pre-trained on large scale datasets is one of the most practical transfer learning paradigm given limited quantity of training samples. To obtain better generalization, using the starting point as the reference (SPAR), either through weights or features, has been successfully applied to transfer learning as a regularizer. However, due to the domain discrepancy between the source and target task, there exists obvious risk of negative transfer in a straightforward manner of knowledge preserving. In this paper, we propose a novel transfer learning algorithm, introducing the idea of Target-awareness REpresentation Disentanglement ( $$\textrm{TRED}$$ TRED ), where the relevant knowledge with respect to the target task is disentangled from the original source model and used as a regularizer during fine-tuning the target model. Two alternative approaches, maximizing Maximum Mean Discrepancy (Max-MMD) and minimizing mutual information (Min-MI) are introduced to achieve the desired disentanglement. Experiments on various real world datasets show that our method stably improves the standard fine-tuning by more than 2% in average. $$\textrm{TRED}$$ TRED also outperforms related state-of-the-art transfer learning regularizers such as $$\mathrm {L^2\text {-}SP}$$ L 2 - SP , $$\textrm{AT}$$ AT , $$\textrm{DELTA}$$ DELTA , and $$\textrm{BSS}$$ BSS . Moreover, our solution is compatible with different choices of disentangling strategies. While the combination of Max-MMD and Min-MI typically achieves higher accuracy, only using Max-MMD can be a preferred choice in applications with low resource budgets.

Molecular property prediction plays a fundamental role in AI-aided drug discovery to identify candidate molecules, which is also essentially a few-shot problem due to lack of labeled data. In this paper, we propose Property-Aware Relation networks (PAR) to handle this problem. We first introduce a property-aware molecular encoder to transform the generic molecular embeddings to property-aware ones. Then, we design a query-dependent relation graph learning module to estimate molecular relation graph and refine molecular embeddings w.r.t. the target property. Thus, the facts that both property-related information and relationships among molecules change across different properties are utilized to better learn and propagate molecular embeddings. Generally, PAR can be regarded as a combination of metric-based and optimization-based few-shot learning method. We further extend PAR to Transferable PAR (T-PAR) to handle the distribution shift, which is common in drug discovery. The keys are joint sampling and relation graph learning schemes, which simultaneously learn molecular embeddings from both source and target domains. Extensive results on benchmark datasets show that PAR and T-PAR consistently outperform existing methods on few-shot and transferable few-shot molecular property prediction tasks, respectively. Besides, ablation and case studies are conducted to validate the rationality of our designs in PAR and T-PAR.

Despite achieving remarkable performance, Federated Learning (FL) encounters two important problems, i.e., low training efficiency and limited computational resources. In this article, we propose a new FL framework, i.e., FedDUMAP, with three original contributions, to leverage the shared insensitive data on the server in addition to the distributed data in edge devices so as to efficiently train a global model. First, we propose a simple dynamic server update algorithm, which takes advantage of the shared insensitive data on the server while dynamically adjusting the update steps on the server in order to speed up the convergence and improve the accuracy. Second, we propose an adaptive optimization method with the dynamic server update algorithm to exploit the global momentum on the server and each local device for superior accuracy. Third, we develop a layer-adaptive model pruning method to carry out specific pruning operations, which is adapted to the diverse features of each layer so as to attain an excellent tradeoff between effectiveness and efficiency. Our proposed FL model, FedDUMAP, combines the three original techniques and has a significantly better performance compared with baseline approaches in terms of efficiency (up to 16.9 times faster), accuracy (up to 20.4% higher), and computational cost (up to 62.6% smaller).

Drug discovery often relies on the successful prediction of protein-ligand binding affinity. Recent advances have shown great promise in applying graph neural networks (GNNs) for better affinity prediction by learning the representations of protein-ligand complexes. However, existing solutions usually treat protein-ligand complexes as topological graph data, thus the 3D geometry-based biomolecular structural information is not fully utilized. The essential intermolecular interactions with long-range dependencies, including type-wise interactions and molecule-wise interactions, are also neglected in GNN models. To this end, we propose a geometry-aware interactive graph neural network (GIaNt) which consists of two components: 3D geometric graph learning network (3DG-Net) and pairwise interactive learning network (Pi-Net). Specifically,3DG-Netiteratively performs the node-edge interaction process to update embeddings of nodes and edges in a unified framework while preserving the 3D geometric factors among atoms, including spatial distance, polar angle and dihedral angle information in 3D space. Moreover,Pi-Netis adopted to incorporate both element type-level and molecule-level interactions. Specially, interactive edges are gathered with a subsequent reconstruction loss to reflect the global type-level interactions. Meanwhile, a pairwise attentive pooling scheme is designed to identify the critical interactive atoms for complex representation learning from a semantic view. An exhaustive experimental study on two benchmarks verifies the superiority ofGIaNt.

While deep learning succeeds in a wide range of tasks, it highly depends on the massive collection of annotated data which is expensive and time-consuming. To lower the cost of data annotation, active learning has been proposed to interactively query an oracle to annotate a small proportion of informative samples in an unlabeled dataset. Inspired by the fact that the samples with higher loss are usually more informative to the model than the samples with lower loss, in this article we present a novel deep active learning approach that queries the oracle for data annotation when the unlabeled sample is believed to incorporate high loss. The core of our approach is a measurement temporal output discrepancy (TOD) that estimates the sample loss by evaluating the discrepancy of outputs given by models at different optimization steps. Our theoretical investigation shows that TOD lower-bounds the accumulated sample loss thus it can be used to select informative unlabeled samples. On basis of TOD, we further develop an effective unlabeled data sampling strategy as well as an unsupervised learning criterion for active learning. Due to the simplicity of TOD, our methods are efficient, flexible, and task-agnostic. Extensive experimental results demonstrate that our approach achieves superior performances than the state-of-the-art active learning methods on image classification and semantic segmentation tasks. In addition, we show that TOD can be utilized to select the best model of potentially the highest testing accuracy from a pool of candidate models.

While deep learning has been widely used for video analytics, such as video classification and action detection, dense action detection with fast-moving subjects from sports videos is still challenging. In this work, we release yet another sports video benchmark P 2 ANet for P ing P ong- A ction detection, which consists of 2,721 video clips collected from the broadcasting videos of professional table tennis matches in World Table Tennis Championships and Olympiads. We work with a crew of table tennis professionals and referees on a specially designed annotation toolbox to obtain fine-grained action labels (in 14 classes) for every ping-pong action that appeared in the dataset, and formulate two sets of action detection problems— action localization and action recognition . We evaluate a number of commonly seen action recognition (e.g., TSM, TSN, Video SwinTransformer, and Slowfast) and action localization models (e.g., BSN, BSN++, BMN, TCANet), using P 2 ANet for both problems, under various settings. These models can only achieve 48% area under the AR-AN curve for localization and 82% top-one accuracy for recognition since the ping-pong actions are dense with fast-moving subjects but broadcasting videos are with only 25 FPS. The results confirm that P 2 ANet is still a challenging task and can be used as a special benchmark for dense action detection from videos. We invite readers to examine our dataset by visiting the following link: https://github.com/Fred1991/P2ANET .

While mislabeled or ambiguously-labeled samples in the training set could negatively affect the performance of deep models, diagnosing the dataset and identifying mislabeled samples helps to improve the generalization power. Training dynamics, i.e., the traces left by iterations of optimization algorithms, have recently been proved to be effective to localize mislabeled samples with hand-crafted features. In this paper, beyond manually designed features, we introduce a novel learning-based solution, leveraging a noise detector, instanced by an LSTM network, which learns to predict whether a sample was mislabeled using the raw training dynamics as input. Specifically, the proposed method trains the noise detector in a supervised manner using the dataset with synthesized label noises and can adapt to various datasets (either naturally or synthesized label-noised) without retraining. We conduct extensive experiments to evaluate the proposed method. We train the noise detector based on the synthesized label-noised CIFAR dataset and test such noise detector on Tiny ImageNet, CUB-200, Caltech-256, WebVision and Clothing1M. Results show that the proposed method precisely detects mislabeled samples on various datasets without further adaptation, and outperforms state-of-the-art methods. Besides, more experiments demonstrate that the mislabel identification can guide a label correction, namely data debugging, providing orthogonal improvements of algorithm-centric state-of-the-art techniques from the data aspect.

Recommender systems usually face confounding bias and selection bias. The former arises when hidden variables determine user/item features and an outcome variable simultaneously while the latter happens due to some biased selection mechanisms, e.g., choosing users based on a specific time or location. How to alleviate such biases has attracted a lot research attention in recent years, but existing approaches mainly focus on one specific source of bias, rather than handle both confounding and selection biases. To this end, we formulate the causal personalized recommendation problem based on the structural causal model (SCM) and a generalization of the notion of backdoor adjustment to account for both biases. Our approach leverages external data of some variables that are also measured without selection bias and uses an adjustment pair based on the derived graphical conditions for identifying conditional causal effects. We present a statistical estimation procedure based on inverse probability weighting to calculate conditional causal effects when training samples are limited. In the presence of confounding and selection biases, we also show how to derive path-specific effects and counterfactual effects, both of which are important for recommendation analysis. We demonstrate the effectiveness of our approach through empirical evaluations.

Denoising is a crucial step for hyperspectral image (HSI) applications. Though witnessing the great power of deep learning, existing HSI denoising methods suffer from limitations in capturing the nonlocal self-similarity. Trans-formers have shown potential in capturing longrange de-pendencies, but few attempts have been made with specifically designed Transformer to model the spatial and spec-tral correlation in HSIs. In this paper, we address these issues by proposing a spectral enhanced rectangle Trans-former, driving it to explore the nonlocal spatial similarity and global spectral low-rank property of HSIs. For the former, we exploit the rectangle self-attention horizontally and vertically to capture the nonlocal similarity in the spatial domain. For the latter, we design a spectral enhancement module that is capable of extracting global underlying low-rank property of spatial-spectral cubes to suppress noise, while enabling the interactions among non-overlapping spatial rectangles. Extensive experiments have been conducted on both synthetic noisy HSIs and real noisy HSIs, showing the effectiveness of our proposed method in terms of both objective metric and subjective visual quality. The code is available at https://github.com/MyuLi/SERT.

Federated learning (FL) is a promising paradigm to enable collaborative model training with decentralized data.However, the training process of Large Language Models (LLMs) generally incurs the update of significant parameters, which limits the applicability of FL techniques to tackle the LLMs in real scenarios.Prompt tuning can significantly reduce the number of parameters to update, but it either incurs performance degradation or low training efficiency.The straightforward utilization of prompt tuning in the FL often raises non-trivial communication costs and dramatically degrades performance.In addition, the decentralized data is generally non-Independent and Identically Distributed (non-IID), which brings client drift problems and thus poor performance.This paper proposes a Parameter-efficient prompt Tuning approach with Adaptive Optimization, i.e., Fed-PepTAO, to enable efficient and effective FL of LLMs.First, an efficient partial prompt tuning approach is proposed to improve performance and efficiency simultaneously.Second, a novel adaptive optimization method is developed to address the client drift problems on both the device and server sides to enhance performance further.Extensive experiments based on 10 datasets demonstrate the superb performance (up to 60.8% in terms of accuracy) and efficiency (up to 97.59% in terms of training time) of FedPepTAO compared with 9 baseline approaches.Our code is available at https://github.com/llm-eff/FedPepTAO.

Fine-tuning pre-trained language models, such as BERT, has shown enormous success among various NLP tasks. Though simple and effective, the process of fine-tuning has been found unstable, which often leads to unexpected poor performance. To increase stability and generalizability, most existing works resort to maintaining the parameters or representations of pre-trained models during fine-tuning. Nevertheless, very little work explores mining the reliable part of pre-learned information that can help to stabilize fine-tuning. To address this challenge, we introduce a novel solution in which we fine-tune BERT with stabilized cross-layer mutual information. Our method aims to preserve the reliable behaviors of cross-layer information propagation, instead of preserving the information itself, of the pre-trained model. Therefore, our method circumvents the domain conflicts between pre-trained and target tasks. We conduct extensive experiments with popular pre-trained BERT variants on NLP datasets, demonstrating the universal effectiveness and robustness of our method.

Long-term time-series forecasting (LTTF) has become a pressing demand in many applications, such as wind power supply planning. Transformer models have been adopted to deliver high prediction capacity because of the high computational self-attention mechanism. Though one could lower the complexity of Transformers by inducing the sparsity in point-wise self-attentions for LTTF, the limited information utilization prohibits the model from exploring the complex dependencies comprehensively. To this end, we propose an efficient Transformer-based model, named Conformer, which differentiates itself from existing methods for LTTF in three aspects: (i) an encoder-decoder architecture incorporating a linear complexity without sacrificing information utilization is proposed on top of sliding-window attention and Stationary and Instant Recurrent Network (SIRN); (ii) a module derived from the normalizing flow is devised to further improve the information utilization by inferring the outputs with the latent variables in SIRN directly; (iii) the inter-series correlation and temporal dynamics in time-series data are modeled explicitly to fuel the downstream self-attention mechanism. Extensive experiments on seven real-world datasets demonstrate that Conformer outperforms the state-of-the-art methods on LTTF and generates reliable prediction results with uncertainty quantification.

Urban villages (UVs) refer to the underdeveloped informal settlement falling behind the rapid urbanization in a city. Since there are high levels of social inequality and social risks in these UVs, it is critical for city managers to discover all UVs for making appropriate renovation policies. Existing approaches to detecting UVs are labor-intensive or have not fully addressed the unique challenges in UV detection such as the scarcity of labeled UVs and the diverse urban patterns in different regions. To this end, we first build an urban region graph (URG) to model the urban area in a hierarchically structured way. Then, we design a novel contextual master-slave framework to effectively detect the urban village from the URG. The core idea of such a framework is to firstly pre-train a basis (or master) model over the URG, and then to adaptively derive specific (or slave) models from the basis model for different regions. The proposed framework can learn to balance the generality and specificity for UV detection in an urban area. Finally, we conduct extensive experiments in three cities to demonstrate the effectiveness of our approach.

Biological Knowledge Graphs (BKGs) can help to model complex biological systems in a structural way to support various tasks. Nevertheless, the incompleteness problem may limit the performance of existing BKGs, which still deserves new methods to reveal the missing relations. Though great efforts have been made to knowledge graph completion, existing methods are not easy to be adapted to the multimodal biological information such as molecular structures and textual descriptions. To this end, we propose a novel co-attention-based multimodal embedding framework, named CamE, for the multimodal BKG completion task. Specifically, we design a Triple Co-Attention (TCA) operator to capture and highlight the same semantic features among different modalities. Based on TCA, we further propose two components to handle multimodal fusion and multimodal entity-relation interaction, respectively. One is the multimodal TCA fusion module to achieve a multimodal joint representation for each entity in the BKG. It aims to project different modal information into a common space by capturing the same semantic features and overcoming the modality gap. The other is the relation-aware interactive TCA module to learn interactive representation by modelling the deep interaction between multimodal entities and relations. Extensive experiments on two real-world multimodal BKG datasets demonstrate that our method significantly outperforms several state-of-the-art baselines, including 10.3% and 16.2% improvement w.r.t MRR and Hits@1 metrics over its best competitors on public DRKG-MM dataset.

Training images with data transformations, e.g., crops, shifts, rotations and color distortions, have been suggested as contrastive examples to evaluate the robustness of deep neural networks against data noises [1]. In this work, we propose a practical framework ContRE (which is the meaning of “against” in French) that uses Contrastive examples for DNN Robustness Estimation. Specifically, ContRE follows the assumption in [2], [3] that robust DNN models with good generalization performance are capable of extracting a consistent set of features and making consistent predictions from the same image under varying data transformations. Incorporating with a set of randomized strategies for well-designed data transformations over the training set, ContREadopts classification errors and Fisher ratios on the generated contrastive examples to assess and analyze the robustness of DNN models, which correlates to the models’ generalization performance. To show the effectiveness and efficiency of ContRE, extensive experiments have been done using various DNN models, e.g., ResNet, VGGNet, DenseNet, EfficientNet, etc., on three open source benchmark datasets, i.e., CIFAR-10, CIFAR-100, and ImageNet, with thorough ablation studies and applicability analyses. Our experiment results confirm that ❨1❩ behaviors of deep models on contrastive examples are strongly correlated to what on the testing set, and ❨2❩ the robustness that ContRE calculates is a robust measure of generalization performance complementing to the testing set in various settings. Codes is to be publicly available.

Federated machine unlearning (FMU) aims to remove the influence of a specified subset of training data upon request from a trained federated learning model. Despite achieving remarkable performance, existing FMU techniques suffer from inefficiency due to two sequential operations of training and retraining/unlearning on large-scale datasets. Our prior study, PCMU, was proposed to improve the efficiency of centralized machine unlearning (CMU) with certified guarantees, by simultaneously executing the training and unlearning operations. This paper proposes a fast FMU algorithm, FFMU, for improving the FMU efficiency while maintaining the unlearning quality. The PCMU method is leveraged to train a local machine learning (MU) model on each edge device. We propose to employ nonlinear functional analysis techniques to refine the local MU models as output functions of a Nemytskii operator. We conduct theoretical analysis to derive that the Nemytskii operator has a global Lipschitz constant, which allows us to bound the difference between two MU models regarding the distance between their gradients. Based on the Nemytskii operator and average smooth local gradients, the global MU model on the server is guaranteed to achieve close performance to each local MU model with the certified guarantees.

The past few years have witnessed an explosive growth of user-generated POI-centric travel blogs, which can provide a comprehensive understanding of a POI for people. However, evaluating the quality of the POI-centric travel blogs and ranking the blogs is not a simple task without domain knowledge or actual travel experience on the target POI. Nevertheless, our insight is that the user search behavior related to the target POI on the online map service can partly valid the rationality of the POIs appearing in the travel blogs, which helps for travel blogs ranking. To this end, in this paper, we propose a novel end-to-end framework for travel blogs ranking, coined Matching POI and Travel Blogs with Multi-view InFormation (MOTIF). Concretely, we first construct two POI graphs as multi-view information: (1) the search-level POI graph which reflects the user behaviors on the online map service; and (2) the document-level POI graph which shows the POI co-occurrence frequency in travel blogs. Then, to better model the intrinsic correlation of the two graphs, we adopt Mutual Information Maximization to align the search-level and document-level semantic spaces. Moreover, we leverage a pair-wise ranking loss for POI-document relevance scoring. Extensive experiments on two real-world datasets demonstrate the superiority of our method.

Making personalized recommendation for cold-start users, who only have a few interaction histories, is a challenging problem in recommendation systems. Recent works leverage hypernetworks to directly map user interaction histories to user-specific parameters, which are then used to modulate predictor by feature-wise linear modulation function. These works obtain the state-of-the-art performance. However, the physical meaning of scaling and shifting in recommendation data is unclear. Instead of using a fixed modulation function and deciding modulation position by expertise, we propose a modulation framework called ColdNAS for user cold-start problem, where we look for proper modulation structure, including function and position, via neural architecture search. We design a search space which covers broad models and theoretically prove that this search space can be transformed to a much smaller space, enabling an efficient and robust one-shot search algorithm. Extensive experimental results on benchmark datasets show that ColdNAS consistently performs the best. We observe that different modulation functions lead to the best performance on different datasets, which validates the necessity of designing a searching-based method. Codes are available at https://github.com/LARS-research/ColdNAS.

MOTIVATION: Protein representation learning methods have shown great potential to many downstream tasks in biological applications. A few recent studies have demonstrated that the self-supervised learning is a promising solution to addressing insufficient labels of proteins, which is a major obstacle to effective protein representation learning. However, existing protein representation learning is usually pretrained on protein sequences without considering the important protein structural information. RESULTS: In this work, we propose a novel structure-aware protein self-supervised learning method to effectively capture structural information of proteins. In particular, a graph neural network model is pretrained to preserve the protein structural information with self-supervised tasks from a pairwise residue distance perspective and a dihedral angle perspective, respectively. Furthermore, we propose to leverage the available protein language model pretrained on protein sequences to enhance the self-supervised learning. Specifically, we identify the relation between the sequential information in the protein language model and the structural information in the specially designed graph neural network model via a novel pseudo bi-level optimization scheme. We conduct experiments on three downstream tasks: the binary classification into membrane/non-membrane proteins, the location classification into 10 cellular compartments, and the enzyme-catalyzed reaction classification into 384 EC numbers, and these experiments verify the effectiveness of our proposed method. AVAILABILITY AND IMPLEMENTATION: The Alphafold2 database is available in https://alphafold.ebi.ac.uk/. The PDB files are available in https://www.rcsb.org/. The downstream tasks are available in https://github.com/phermosilla/IEConv\_proteins/tree/master/Datasets. The code of the proposed method is available in https://github.com/GGchen1997/STEPS_Bioinformatics.

Abstract Although more layers and more parameters generally improve the accuracy of the models, such big models generally have high computational complexity and require big memory, which exceed the capacity of small devices for inference and incurs long training time. In addition, it is difficult to afford long training time and inference time of big models even in high performance servers, as well. As an efficient approach to compress a large deep model (a teacher model) to a compact model (a student model), knowledge distillation emerges as a promising approach to deal with the big models. Existing knowledge distillation methods cannot exploit the elastic available computing resources and correspond to low efficiency. In this paper, we propose an Elastic Deep Learning framework for knowledge Distillation, that is, EDL‐Dist. The advantages of EDL‐Dist are threefold. First, the inference and the training process is separated. Second, elastic available computing resources can be utilized to improve the efficiency. Third, fault‐tolerance of the training and inference processes is supported. We take extensive experimentation to show that the throughput of EDL‐Dist is up to 3.125 times faster than the baseline method (online knowledge distillation) while the accuracy is similar or higher.

Summary In recent years, data are typically distributed in multiple organizations while the data security is becoming increasingly important. Federated learning (FL), which enables multiple parties to collaboratively train a model without exchanging the raw data, has attracted more and more attention. Based on the distribution of data, FL can be realized in three scenarios, that is, horizontal, vertical, and hybrid. In this article, we propose to combine distributed machine learning techniques with vertical FL and propose a distributed vertical federated learning (DVFL) approach. The DVFL approach exploits a fully distributed architecture within each party in order to accelerate the training process. In addition, we exploit homomorphic encryption to protect the data against honest‐but‐curious participants. We conduct extensive experimentation in a large‐scale cluster environment and a cloud environment in order to show the efficiency and scalability of our proposed approach. The experiments demonstrate the good scalability of our approach and the significant efficiency advantage (up to 6.8 times with a single server and 15.1 times with multiple servers in terms of the training time) compared with baseline frameworks.

Due to privacy concerns of users and law enforcement in data security and privacy, it becomes more and more difficult to share data among organizations. Data federation brings new opportunities to the data-related cooperation among organizations by providing abstract data interfaces. With the development of cloud computing, organizations store data on the cloud to achieve elasticity and scalability for data processing. The existing data placement approaches generally only consider one aspect, which is either execution time or monetary cost, and do not consider data partitioning for hard constraints. In this paper, we propose an approach to enable data processing on the cloud with the data from different organizations. The approach consists of a data federation platform namedFedCubeand a Lyapunov-based data placement algorithm.FedCubeenables data processing on the cloud. We use the data placement algorithm to create a plan in order to partition and store data on the cloud so as to achieve multiple objectives while satisfying the constraints based on a multi-objective cost model. The cost model is composed of two objectives, i.e., reducing monetary cost and execution time. We present an experimental evaluation to show our proposed algorithm significantly reduces the total cost (up to 69.8%) compared with existing approaches.

Whilelearning to rank(LTR) has been widely used in web search to prioritize most relevant webpages among the retrieved contents subject to the input queries, the traditional LTR models fail to deliver decent performance due to two main reasons: 1) the lack of well-annotated query-webpage pairs with ranking scores to cover search queries of various popularity, and 2) ill-trained models based on a limited number of training samples with poor generalization performance. To improve the performance of LTR models, tremendous efforts have been done from above two aspects, such as enlarging training sets with pseudo-labels of ranking scores by self-training, or refining the features used for LTR through feature extraction and dimension reduction. Though LTR performance has been marginally increased, we still believe these methods could be further improved in the newly-fashioned “interpolating regime”. Specifically, instead of lowering the number of features used for LTR models, our work proposes to transform original data with random Fourier feature, so as to over-parameterize the downstream LTR models (e.g., GBRank or LightGBM) with features in ultra-high dimensionality and achieve superb generalization performance. Furthermore, rather than self-training with pseudo-labels produced by the same LTR model in a “self-tuned” fashion, the proposed method incorporates the diversity of prediction results between the listwise and pointwise LTR models while co-training both models with a cyclic labeling-prediction pipeline in a “ping-pong” manner. We deploy the proposedCo-trained andOver-parameterizedLTRsystemCOLTRat Baidu search and evaluateCOLTRwith a large number of baseline methods. The results show thatCOLTRcould achieve$\Delta NDCG_{4}$= 3.64%$\sim$4.92%, compared to baselines, under various ratios of labeled samples. We also conduct a 7-day A/B Test using the realistic web traffics of Baidu Search, where we can still observe significant performance improvement around$\Delta NDCG_{4}$= 0.17%$\sim$0.92% in real-world applications.COLTRperforms consistently both in online and offline experiments.

While spatial-temporal environment monitoring has become an indispensable way to collect data for enabling smart cities and intelligent transportation applications, the cost to deploy, operate and maintain a sensor network with sensors and massive communication infrastructure is too high to bear. Compared to the infrastructure-based sensing approach, community sensing, or namely mobile crowdsensing, that leverage community members' mobile devices to collect data becomes a feasible way to scale up the spatial-temporal coverage of the sensing system. However, a community sensing system would need to aggregate sensors and location data from community members and thus would raise concerns on privacy and data security In this paper, we present a novel community sensing paradigm AFCS -Sensor and Location Data Aggregation-Free Community Sensing, which is designed to obtain the environment information (e.g., spatial-temporal distributions of air pollution, temperature, and bike-shares) in each subarea of the target area, without aggregating sensor and location data collected by community members. AFCS proposes to orchestrate with the Trusted Execution Environments (TEEs) of every community member's mobile device to cover the communication, computation and storage with spatial-temporal data. Further, AFCS proposes a novel Decentralized Spatial-Temporal Compressive Sensing framework based on Parallelized Stochastic Gradient Descent. Through learning the latent structure of the spatial-temporal data via decentralized optimization, AFCS approximates the value of the sensor data in each subarea (both covered and uncovered) for each sensing cycle using the sensor data locally stored in every member's TEE instance. Experiments based on real-world datasets and the Virtual Mobile Infrastructure (VMI) with TEE emulations demonstrate that AFCS exhibits low approximation error (i.e., less than 0:2°C in city-wide temperature sensing, 10 units of PM2.5 index in urban air pollution sensing, and 2 bikes in city-wide bike sharing prediction) and performs comparably to (sometimes better than) state-of-the-art algorithms based on the data aggregation and centralized computation

To understand human behaviors, action recognition based on videos is a common approach. Compared with image-based action recognition, videos provide much more information, reducing the ambiguity of actions. In the last decade, many works focus on datasets, novel models and learning approaches have improved video action recognition to a higher level. However, there are challenges and unsolved problems, in particular in sports analytics where data collection and labeling are more sophisticated, requiring people with domain knowledge and even sport professionals to annotate data. In addition, the actions could be extremely fast and it becomes difficult to recognize them. Moreover, in team sports like football and basketball, one action could involve multiple players, and to correctly recognize them, we need to analyze all players, which is relatively complicated. In this paper, we present a survey on video action recognition for sports analytics. We introduce more than ten types of sports, including team sports, such as football, basketball, volleyball, hockey and individual sports, such as figure skating, gymnastics, table tennis, tennis, diving and badminton. Then we compare numerous existing frameworks for sports analysis to present status quo of video action recognition in both team sports and individual sports. Finally, we discuss the challenges and unsolved problems in this area and to facilitate sports analytics, we develop a toolbox using PaddlePaddle, which supports football, basketball, table tennis and figure skating action recognition.

Recent years have witnessed a large amount of decentralized data in various (edge) devices of end-users, while the decentralized data aggregation remains complicated for machine learning jobs because of regulations and laws. As a practical approach to handling decentralized data, Federated Learning (FL) enables collaborative global machine learning model training without sharing sensitive raw data. The servers schedule devices to jobs within the training process of FL. In contrast, device scheduling with multiple jobs in FL remains a critical and open problem. In this article, we propose a novel multi-job FL framework, which enables the training process of multiple jobs in parallel. The multi-job FL framework is composed of a system model and a scheduling method. The system model enables a parallel training process of multiple jobs, with a cost model based on the data fairness and the training time of diverse devices during the parallel training process. We propose a novel intelligent scheduling approach based on multiple scheduling methods, including an original reinforcement learning-based scheduling method and an original Bayesian optimization-based scheduling method, which corresponds to a small cost while scheduling devices to multiple jobs. We conduct extensive experimentation with diverse jobs and datasets. The experimental results reveal that our proposed approaches significantly outperform baseline approaches in terms of training time (up to 12.73 times faster) and accuracy (up to 46.4% higher).

While recommender systems have been ubiquitously used in digital marketing and online business development, the conversions of online advertising for mobile apps installation and activation sometimes are far from satisfactory, due to the lack of feedback from App-related activities, leading to a poor record of Return on Investment (RoI). Though the advertisers, e.g., App operators and App Store, are granted to log users’ app-related activities such as installation, activation, usages, and preferences per the agreement, they usually limit the access to such data from advertisement publishers, due to the privacy concerns. To improve conversions of online advertising under privacy controls, we proposeFeynman—afederated learning-based advertising platform for ecosystems-orientedmobileapps recommendation.Feynmanaims at improving the RoI of mobile app recommendation from an ecosystem's perspective, i.e., per investment in advertising an app (Goal. 1) increasing the number of new installs/users of the app, and then (Goal. 2) increasing the number of new active users (preferably with frequent in-app purchase activities). Incorporating with a federated computing platform,Feynmanleverages users’ records stored in advertisers to refine the pool of targeting users for ads distribution, and jointly builds the predictive models for users’ purchase activities forecasting using features from the Ads publisher and the advertiser. With refined target pools and more accurate models,Feynmanhas successfully helped several mobile apps in China by attracting more than 100 million users to further enlarge their user populations and revenues from in-app purchases. Note that rather than proposing new techniques for federated learning, the design ofFeynmandedicates to show its promising performance in the industrial practices of advertising using federated computing and privacy protected strategies. In three cases that we report in this paper,Feynmanoutperforms the state-of-the-art plans in terms of several key measurements, including Click-Through Rates (CTR), Conversion Rate (CVR), Cost per Action (CPA), and Non-targeting User Hit-Rates (NTHR).

Recently many efforts have been devoted to applying graph neural networks (GNNs) to molecular property prediction which is a fundamental task for computational drug and material discovery. One of major obstacles to hinder the successful prediction of molecular property by GNNs is the scarcity of labeled data. Though graph contrastive learning (GCL) methods have achieved extraordinary performance with insufficient labeled data, most focused on designing data augmentation schemes for general graphs. However, the fundamental property of a molecule could be altered with the augmentation method (like random perturbation) on molecular graphs. Whereas, the critical geometric information of molecules remains rarely explored under the current GNN and GCL architectures. To this end, we propose a novel graph contrastive learning method utilizing the geometry of the molecule across 2D and 3D views, which is named GeomGCL. Specifically, we first devise a dual-view geometric message passing network (GeomMPNN) to adaptively leverage the rich information of both 2D and 3D graphs of a molecule. The incorporation of geometric properties at different levels can greatly facilitate the molecular representation learning. Then a novel geometric graph contrastive scheme is designed to make both geometric views collaboratively supervise each other to improve the generalization ability of GeomMPNN. We evaluate GeomGCL on various downstream property prediction tasks via a finetune process. Experimental results on seven real-life molecular datasets demonstrate the effectiveness of our proposed GeomGCL against state-of-the-art baselines.

Recent years have witnessed a large amount of decentralized data in multiple (edge) devices of end-users, while the aggregation of the decentralized data remains difficult for machine learning jobs due to laws or regulations. Federated Learning (FL) emerges as an effective approach to handling decentralized data without sharing the sensitive raw data, while collaboratively training global machine learning models. The servers in FL need to select (and schedule) devices during the training process. However, the scheduling of devices for multiple jobs with FL remains a critical and open problem. In this paper, we propose a novel multi-job FL framework to enable the parallel training process of multiple jobs. The framework consists of a system model and two scheduling methods. In the system model, we propose a parallel training process of multiple jobs, and construct a cost model based on the training time and the data fairness of various devices during the training process of diverse jobs. We propose a reinforcement learning-based method and a Bayesian optimization-based method to schedule devices for multiple jobs while minimizing the cost. We conduct extensive experimentation with multiple jobs and datasets. The experimental results show that our proposed approaches significantly outperform baseline approaches in terms of training time (up to 8.67 times faster) and accuracy (up to 44.6% higher).

Although remarkable progress on the neural table-to-text methods has been made, the generalization issues hinder the applicability of these models due to the limited source tables.Large-scale pretrained language models sound like a promising solution to tackle such issues.However, how to effectively bridge the gap between the structured table and the text input by fully leveraging table information to fuel the pretrained model is still not well explored.Besides, another challenge of integrating the deliberation mechanism into the textto-text pretrained model for solving the tableto-text task remains seldom studied.In this paper, to implement the table-to-text generation with pretrained language model, we propose a table structure understanding and text deliberating approach, namely TASD.To be specific, we devise a three-layered multi-head attention network to realize the table-structureaware text generation model with the help of the pretrained language model.Furthermore, a multi-pass decoder framework is adopted to enhance the capability of polishing generated text for table descriptions.The empirical studies, as well as human evaluation, on two public datasets, validate that our approach can generate faithful and fluent descriptive texts for different types of tables.

Short text classification (STC) is hard as short texts lack context information and labeled data is not enough.Graph neural networks obtain the state-of-the-art on STC since they can merge various auxiliary information via the message passing framework.However, existing works conduct transductive learning, which requires retraining to accommodate new samples and takes large memory.In this paper, we present SimpleSTC which handles inductive STC problem but only leverages words.We construct word graph from an external large corpus to compensate for the lack of semantic information, and learn text graph to handle the lack of labeled data.Results show that Sim-pleSTC obtains state-of-the-art performance with lower memory consumption and faster inference speed. 1

Arbitrary image style transfer is a challenging task which aims to stylize a content image conditioned on arbitrary style images. In this task the feature-level content-style transformation plays a vital role for proper fusion of features. Existing feature transformation algorithms often suffer from loss of content or style details, non-natural stroke patterns, and unstable training. To mitigate these issues, this paper proposes a new feature-level style transformation technique, named Style Projection, for parameter-free, fast, and effective content-style transformation. This paper further presents a real-time feed-forward model to leverage Style Projection for arbitrary image style transfer, which includes a regularization term for matching the semantics between input contents and stylized outputs. Extensive qualitative analysis, quantitative evaluation, and user study have demonstrated the effectiveness and efficiency of the proposed methods.

Recent studies on federated learning (FL) have sought to solve the system heterogeneity issue by designing customized local models for different clients. However, public dataset introduction, sensitive information exchange, non-trivial computational cost, or particular architecture requirement limit the applicability of most of them in real scenarios. This paper presents a novel federated fingerprint learning model for making full use of the computing power of each client with the customized local models for improving the FL convergence, while keeping the data and sensitive information safe and local. First, we decompose the parameters of each local model into two types of parameters: rigid ones that have fixed model architecture for ensuring the convergence of global model training and elastic ones that contain customized model structure and size for allowing to make full use of the computing power of each client based on individual data scale. Second, we adopt the standard FL scheme to update and aggregate the local rigid parameters. We introduce a Gaussian distribution as auxiliary input and output K local fingerprints respectively for the elastic parameters of all K local models. The server aggregates K local fingerprints into a global one and sends it back to the clients. A fingerprint-based aggregation strategy makes the local models indirectly receive the aggregated elastic parameters through the aggregation of K local fingerprints while fixing data locally. Last but not least, we design a parameter masking method to mask the rigid parameters irrelevant to the local classification task in the local models. We develop a parameter separation method to guarantee that the combination of unmasked rigid parameters in all local models are able to cover all the rigid parameters as many as possible, for further raising the utilization rate of each rigid parameter.

The federated learning (FL) framework enables edge clients to collaboratively learn a shared inference model while keeping privacy of training data on clients. Recently, many heuristics efforts have been made to generalize centralized adaptive optimization methods, such as SGDM, Adam, AdaGrad, etc., to federated settings for improving convergence and accuracy. However, there is still a paucity of theoretical principles on where to and how to design and utilize adaptive optimization methods in federated settings. This work aims to develop novel adaptive optimization methods for FL from the perspective of dynamics of ordinary differential equations (ODEs). First, an analytic framework is established to build a connection between federated optimization methods and decompositions of ODEs of corresponding centralized optimizers. Second, based on this analytic framework, a momentum decoupling adaptive optimization method, FedDA, is developed to fully utilize the global momentum on each local iteration and accelerate the training convergence. Last but not least, full batch gradients are utilized to mimic centralized optimization in the end of the training process to ensure the convergence and overcome the possible inconsistency caused by adaptive optimization methods.

Federated Learning (FL) enables training a global model without sharing the decentralized raw data stored on multiple devices to protect data privacy. Due to the diverse capacity of the devices, FL frameworks struggle to tackle the problems of straggler effects and outdated models. In addition, the data heterogeneity incurs severe accuracy degradation of the global model in the FL training process. To address aforementioned issues, we propose a hierarchical synchronous FL framework, i.e., FedHiSyn. FedHiSyn first clusters all available devices into a small number of categories based on their computing capacity. After a certain interval of local training, the models trained in different categories are simultaneously uploaded to a central server. Within a single category, the devices communicate the local updated model weights to each other based on a ring topology. As the efficiency of training in the ring topology prefers devices with homogeneous resources, the classification based on the computing capacity mitigates the impact of straggler effects. Besides, the combination of the synchronous update of multiple categories and the device communication within a single category help address the data heterogeneity issue while achieving high accuracy. We evaluate the proposed framework based on MNIST, EMNIST, CIFAR10 and CIFAR100 datasets and diverse heterogeneous settings of devices. Experimental results show that FedHiSyn outperforms six baseline methods, e.g., FedAvg, SCAFFOLD, and FedAT, in terms of training accuracy and efficiency.

Despite achieving remarkable performance, Federated Learning (FL) suffers from two critical challenges, i.e., limited computational resources and low training efficiency. In this paper, we propose a novel FL framework, i.e., FedDUAP, with two original contributions, to exploit the insensitive data on the server and the decentralized data in edge devices to further improve the training efficiency. First, a dynamic server update algorithm is designed to exploit the insensitive data on the server, in order to dynamically determine the optimal steps of the server update for improving the convergence and accuracy of the global model. Second, a layer-adaptive model pruning method is developed to perform unique pruning operations adapted to the different dimensions and importance of multiple layers, to achieve a good balance between efficiency and effectiveness. By integrating the two original techniques together, our proposed FL model, FedDUAP, significantly outperforms baseline approaches in terms of accuracy (up to 4.8% higher), efficiency (up to 2.8 times faster), and computational cost (up to 61.9% smaller).

Trip planning, which targets at planning a trip consisting of several ordered Points of Interest (POIs) under user-provided constraints, has long been treated as an important application for location-based services. The goal of trip planning is to maximize the chance that the users will follow the planned trip while it is difficult to directly quantify and optimize the chance. Conventional methods either leverage statistical analysis to rank POIs to form a trip or generate trips following pre-defined objectives based on constraint programming to bypass such a problem. However, these methods may fail to reflect the complex latent patterns hidden in the human mobility data. On the other hand, though there are a few deep learning-based trip recommendation methods, these methods still cannot handle the time budget constraint so far. To this end, we propose a TIme-aware Neural Trip Planning (TINT) framework to tackle the above challenges. First of all, we devise a novel attention-based encoder-decoder trip generator that can learn the correlations among POIs and generate trips under given constraints. Then, we propose a specially-designed reinforcement learning (RL) paradigm to directly optimize the objective to obtain an optimal trip generator. For this purpose, we introduce a discriminator, which distinguishes the generated trips from real-life trips taken by users, to provide reward signals to optimize the generator. Subsequently, to ensure the feedback from the discriminator is always instructive, we integrate an adversarial learning strategy into the RL paradigm to update the trip generator and the discriminator alternately. Moreover, we devise a novel pre-training schema to speed up the convergence for an efficient training process. Extensive experiments on four real-world datasets validate the effectiveness and efficiency of our framework, which shows that TINT could remarkably outperform the state-of-the-art baselines within short response time.

In this paper, we propose a learning-based moving-object tracking method utilizing the newly developed LiDAR sensor, Frequency Modulated Continuous Wave (FMCW) LiDAR. Compared with most existing commercial LiDAR sensors, FMCW LiDAR can provide additional Doppler velocity information to each 3D point of the point clouds. Benefiting from this, we can generate instance labels as ground truth in a semi-automatic manner. Given the labels, we propose a contrastive learning framework, which pulls together the features from the same instance in embedding space and pushes apart the features from different instances, to improve the tracking quality. Extensive experiments are conducted on the recorded driving data, and the results show that our method outperforms the baseline methods by a large margin.

Time series forecasting has been a widely explored task of great importance in many applications. However, it is common that real-world time series data are recorded in a short time period, which results in a big gap between the deep model and the limited and noisy time series. In this work, we propose to address the time series forecasting problem with generative modeling and propose a bidirectional variational auto-encoder (BVAE) equipped with diffusion, denoise, and disentanglement, namely D3VAE. Specifically, a coupled diffusion probabilistic model is proposed to augment the time series data without increasing the aleatoric uncertainty and implement a more tractable inference process with BVAE. To ensure the generated series move toward the true target, we further propose to adapt and integrate the multiscale denoising score matching into the diffusion process for time series forecasting. In addition, to enhance the interpretability and stability of the prediction, we treat the latent variable in a multivariate manner and disentangle them on top of minimizing total correlation. Extensive experiments on synthetic and real-world data show that D3VAE outperforms competitive algorithms with remarkable margins. Our implementation is available at https://github.com/PaddlePaddle/PaddleSpatial/tree/main/research/D3VAE.

Given an unsupervised novelty detection task on a new dataset, how can we automatically select a ''best'' detection model while simultaneously controlling the error rate of the best model? For novelty detection analysis, numerous detectors have been proposed to detect outliers on a new unseen dataset based on a score function trained on available clean data. However, due to the absence of labeled data for model evaluation and comparison, there is a lack of systematic approaches that are able to select a ''best'' model/detector (i.e., the algorithm as well as its hyperparameters) and achieve certain error rate control simultaneously. In this paper, we introduce a unified data-driven procedure to address this issue. The key idea is to maximize the number of detected outliers while controlling the false discovery rate (FDR) with the help of Jackknife prediction. We establish non-asymptotic bounds for the false discovery proportions and show that the proposed procedure yields valid FDR control under some mild conditions. Numerical experiments on both synthetic and real data validate the theoretical results and demonstrate the effectiveness of our proposed AutoMS method. The code is available at https://github.com/ZhangYifan1996/AutoMS.

Online healthcare services can provide unlimited and in-time medical information to users, which promotes social goods and breaks the barriers of locations. However, understanding the user intents behind the medical related queries is a challenging problem. Medical search queries are usually short and noisy, lack strict syntactic structure, and also require professional background to understand the medical terms. The medical intents are fine-grained, making them hard to recognize. In addition, many intents only have a few labeled data. To handle these problems, we propose a few-shot learning method for medical search query intent recognition called MEDIC. We extract co-click queries from user search logs as weak supervision to compensate for the lack of labeled data. We also design a new query encoder which learns to represent queries as a combination of semantic knowledge recorded in an external medical knowledge graph, syntactic knowledge which marks the grammatical role of each word in the query, and generic knowledge which is captured by language models pretrained from large-scale text corpus. Experimental results on a real medical search query intent recognition dataset validate the effectiveness of MEDIC.

Pre-travel out-of-town recommendation aims to recommend Point-of-Interests (POIs) to the users who plan to travel out of their hometown in the near future yet have not decided where to go, i.e., their destination regions and POIs both remain unknown. It is a non-trivial task since the searching space is vast, which may lead to distinct travel experiences in different out-of-town regions and eventually confuse decision-making. Besides, users' out-of-town travel behaviors are affected not only by their personalized preferences but heavily by others' travel behaviors. To this end, we propose a Crowd-Aware Pre-Travel Out-of-town Recommendation framework (CAPTOR) consisting of two major modules: spatial-affined conditional random field (SA-CRF) and crowd behavior memory network (CBMN). Specifically, SA-CRF captures the spatial affinity among POIs while preserving the inherent information of POIs. Then, CBMN is proposed to maintain the crowd travel behaviors w.r.t. each region through three affiliated blocks reading and writing the memory adaptively. We devise the elaborated metric space with a dynamic mapping mechanism, where the users and POIs are distinguishable both inherently and geographically. Extensive experiments on two real-world nationwide datasets validate the effectiveness of CAPTOR against the pre-travel out-of-town recommendation task.

Over the last decade, machine learning (ML) and deep learning (DL) algorithms have significantly evolved and been employed in diverse applications, such as computer vision, natural language processing, automated speech recognition, etc. Real-time safety-critical embedded and Internet of Things (IoT) systems, such as autonomous driving systems, UAVs, drones, security robots, etc., heavily rely on ML/DL-based technologies, accelerated with the improvement of hardware technologies. The cost of a deadline (required time constraint) missed by ML/DL algorithms would be catastrophic in these safety-critical systems. However, ML/DL algorithm-based applications have more concerns about accuracy than strict time requirements. Accordingly, researchers from the real-time systems (RTSs) community address the strict timing requirements of ML/DL technologies to include in RTSs. This article will rigorously explore the state-of-the-art results emphasizing the strengths and weaknesses in ML/DL-based scheduling techniques, accuracy versus execution time tradeoff policies of ML algorithms, and security and privacy of learning-based algorithms in real-time IoT systems.

Mobile sensing systems have been widely used as a practical approach to collect behavioral and health-related information from individuals and to provide timely intervention to promote health and well being, such as mental health and chronic care. As the objectives of mobile sensing could be eitherpersonalized medicine for individualsorpublic health for populations, in this work, we review the design of these mobile sensing systems, and propose to categorize the design of these systems in two paradigms—1)personal sensingand 2)crowdsensingparadigms. While both sensing paradigms might incorporate common ubiquitous sensing technologies, such aswearable sensors,mobility monitoring,mobile data offloading, andcloud-based data analyticsto collect and process sensing data from individuals, we present two novel taxonomy systems based on the: 1)sensing objectives(e.g., goals of mobile health (mHealth) sensing systems and how technologies achieve the goals) and 2)the sensing systems design and implementation (D&I)(e.g., designs of mHealth sensing systems and how technologies are implemented). With respect to the two paradigms and two taxonomy systems, this work systematically reviews this field. Specifically, we first present technical reviews on the mHealth sensing systems in eight common/popular healthcare issues, ranging from depression and anxiety to COVID-19. By summarizing the mHealth sensing systems, we comprehensively survey the research works using the two taxonomy systems, where we systematically review thesensing objectivesandsensing systems D&Iwhile mapping the related research works onto the life-cycles of mHealth Sensing, i.e.: 1)sensing task creation and participation; 2)(health surveillance and data collection; and 3)data analysis and knowledge discovery. In addition to summarization, the proposed taxonomy systems also help the potential directions of mobile sensing for health from both personalized medicine and population health perspectives. Finally, we attempt to test and discuss the validity of our scientific approaches to the survey.

Recent advances in wireless sensing techniques have made it possible to recognize hand gestures using channel state information (CSI) in commodity WiFi devices. Existing WiFi-based gesture recognition systems mainly use learning-based pattern recognition methods to recognize different gestures, however, these methods fail to work well when the locations of transceivers, the relative location and orientation of the hand with respect to transceivers, and/or the hand gesturing size change, leading to inconsistent signal patterns caused by those factors. Although some recent efforts have been made to address the so-called “domain-dependent” gesture recognition problem, they either require prior knowledge on initial locations of the hand and WiFi devices or need to train several classifiers for the specific domains. Different from the state-of-the-art methods, we construct two distinct features from a hand-oriented view (rather than from a transceiver’s view), namely, the dynamic phase vector (DPV) and motion rotation variable (MRV), which are quite consistent in characterizing a big set of handwriting gestures, despite significant change in locations of transceivers, the relative location and orientation of the hand with respect to transceivers, and the drawing sizes. We further incorporate a hierarchical sensing framework and develop HandGest—a real-time handwriting gesture recognition system using commodity WiFi devices, to precisely recognize a great number of “in-the-air” handwritings based on the aforementioned two domain-independent features and a pipeline of specific features. Extensive experiments have been done in practical settings with 20 volunteers, evaluation results demonstrate that HandGest outperforms state-of-the-art methods on a large number of handwritings with different transceivers’ location, different initial hand locations and orientations, as well as different drawing sizes. Given its superior performance, we believe that HandGest paves a new way to enhance the real-world practicality of WiFi-based gesture recognition.

Techniques to explain the predictions of deep neural networks (DNNs) have been largely required for gaining insights into the black boxes. We introduce InterpretDL, a toolkit of explanation algorithms based on PaddlePaddle, with uniformed programming interfaces and "plug-and-play" designs. A few lines of codes are needed to obtain the explanation results without modifying the structure of the model. InterpretDL currently contains 16 algorithms, explaining training phases, datasets, global and local behaviors of post-trained deep models. InterpretDL also provides a number of tutorial examples and showcases to demonstrate the capability of InterpretDL working on a wide range of deep learning models, e.g., Convolutional Neural Networks (CNNs), Multi-Layer Preceptors (MLPs), Transformers, etc., for various tasks in both Computer Vision (CV) and Natural Language Processing (NLP). Furthermore, InterpretDL modularizes the implementations, making efforts to support the compatibility across frameworks. The project is available at https://github.com/PaddlePaddle/InterpretDL.

Transfer learning through fine-tuning a pre-trained neural network with an extremely large dataset, such as ImageNet, can significantly improve and accelerate training while the accuracy is frequently bottlenecked by the limited dataset size of the new target task. To solve the problem, some regularization methods, constraining the outer layer weights of the target network using the starting point as references (SPAR), have been studied. In this article, we propose a novel regularized transfer learning framework \operatorname{DELTA} , namely DE ep L earning T ransfer using Feature Map with A ttention . Instead of constraining the weights of neural network, \operatorname{DELTA} aims at preserving the outer layer outputs of the source network. Specifically, in addition to minimizing the empirical loss, \operatorname{DELTA} aligns the outer layer outputs of two networks, through constraining a subset of feature maps that are precisely selected by attention that has been learned in a supervised learning manner. We evaluate \operatorname{DELTA} with the state-of-the-art algorithms, including L^2 and \emph {L}^2\text{-}SP . The experiment results show that our method outperforms these baselines with higher accuracy for new tasks. Code has been made publicly available. 1

Data science competitions are becoming increasingly popular for enterprises collecting advanced innovative solutions and allowing contestants to sharpen their data science skills. Most existing studies about data science competitions have a focus on improving task-specific data science techniques, such as algorithm design and parameter tuning. However, little effort has been made to understand the data science competition itself. To this end, in this article, we shed light on the team’s competition performance, and investigate the team’s evolving performance in the crowd-sourcing competitive innovation context. Specifically, we first acquire and construct multi-sourced datasets of various data science competitions, including the KDD Cup 2019 machine learning competition and beyond. Then, we conduct an empirical analysis to identify and quantify a rich set of features that are significantly correlated with teams’ future performances. By leveraging team’s rank as a proxy, we observe “the stronger, the stronger” rule; that is, top-ranked teams tend to keep their advantages and dominate weaker teams for the rest of the competition. Our results also confirm that teams with diversified backgrounds tend to achieve better performances. After that, we formulate the team’s future rank prediction problem and propose the Multi-Task Representation Learning (MTRL) framework to model both static features and dynamic features. Extensive experimental results on four real-world data science competitions demonstrate the team’s future performance can be well predicted by using MTRL. Finally, we envision our study will not only help competition organizers to understand the competition in a better way, but also provide strategic implications to contestants, such as guiding the team formation and designing the submission strategy.

Network alignment, which aims at learning a matching between the same entities across multiple information networks, often suffers challenges from feature inconsistency, high-dimensional features, to unstable alignment results. This article presents a novel network alignment framework, Unsupervised Adversarial learning based Network Alignment(UANA), that combines generative adversarial network (GAN) and reinforcement learning (RL) techniques to tackle the above critical challenges. First, we propose a bidirectional adversarial network distribution matching model to perform the bidirectional cross-network alignment translations between two networks, such that the distributions of real and translated networks completely overlap together. In addition, two cross-network alignment translation cycles are constructed for training the unsupervised alignment without the need of prior alignment knowledge. Second, in order to address the feature inconsistency issue, we integrate a dual adversarial autoencoder module with an adversarial binary classification model together to project two copies of the same vertices with high-dimensional inconsistent features into the same low-dimensional embedding space. This facilitates the translations of the distributions of two networks in the adversarial network distribution matching model. Finally, we develop an RL based optimization approach to solve the vertex matching problem in the discrete space of the GAN model, i.e., directly select the vertices in target networks most relevant to the vertices in source networks, without unstable similarity computation that is sensitive to discriminative features and similarity metrics. Extensive evaluation on real-world graph datasets demonstrates the outstanding capability of UANA to address the unsupervised network alignment problem, in terms of both effectiveness and scalability.

Temporal relational modeling in video is essential for human action understanding, such as action recognition and action segmentation. Although Graph Convolution Networks (GCNs) have shown promising advantages in relation reasoning on many tasks, it is still a challenge to apply graph convolution networks on long video sequences effectively. The main reason is that large number of nodes (i.e., video frames) makes GCNs hard to capture and model temporal relations in videos. To tackle this problem, in this paper, we introduce an effective GCN module, Dilated Temporal Graph Reasoning Module (DTGRM), designed to model temporal relations and dependencies between video frames at various time spans. In particular, we capture and model temporal relations via constructing multi-level dilated temporal graphs where the nodes represent frames from different moments in video. Moreover, to enhance temporal reasoning ability of the proposed model, an auxiliary self-supervised task is proposed to encourage the dilated temporal graph reasoning module to find and correct wrong temporal relations in videos. Our DTGRM model outperforms state-of-the-art action segmentation models on three challenging datasets: 50Salads, Georgia Tech Egocentric Activities (GTEA), and the Breakfast dataset. The code is available at https://github.com/redwang/DTGRM.

Accurate and timely air quality and weather predictions are of great importance to urban governance and human livelihood. Though many efforts have been made for air quality or weather prediction, most of them simply employ one another as feature input, which ignores the inner-connection between two predictive tasks. On the one hand, the accurate prediction of one task can help improve another task's performance. On the other hand, geospatially distributed air quality and weather monitoring stations provide additional hints for city-wide spatiotemporal dependency modeling. Inspired by the above two insights, in this paper, we propose the Multi-adversarial spatiotemporal recurrent Graph Neural Networks (MasterGNN) for joint air quality and weather prediction. Specifically, we first propose a heterogeneous recurrent graph neural network to model the spatiotemporal autocorrelation among air quality and weather monitoring stations. Then, we develop a multi-adversarial graph learning framework to against observation noise propagation introduced by spatiotemporal modeling. Moreover, we introduce an adaptive training strategy by formulating multi-adversarial learning as a multi-task learning problem. Finally, extensive experiments on two real-world datasets show that MasterGNN achieves the best performance compared with seven baselines on both air quality and weather prediction tasks.

Transportation demand prediction is of great importance to urban governance and has become an essential function in many online applications. While many efforts have been made for regional transportation demand prediction, predicting the diversified transportation demand for different communities (e.g., the aged, the juveniles) remains an unexplored problem. However, this task is challenging because of the joint influence of spatio-temporal correlation among regions and implicit correlation among different communities. To this end, in this paper, we propose the Multi-task Spatio-Temporal Network with Mutually-supervised Adaptive task grouping (Ada-MSTNet) for community-aware transportation demand prediction. Specifically, we first construct a sequence of multi-view graphs from both spatial and community perspectives, and devise a spatio-temporal neural network to simultaneously capture the sophisticated correlations between regions and communities, respectively. Then, we propose an adaptively clustered multi-task learning module, where the prediction of each region-community specific transportation demand is regarded as distinct task. Moreover, a mutually supervised adaptive task grouping strategy is introduced to softly cluster each task into different task groups, by leveraging the supervision signal from one another graph view. In such a way, Ada-MSTNet is not only able to share common knowledge among highly related communities and regions, but also shield the noise from unrelated tasks in an end-to-end fashion. Finally, extensive experiments on two real-world datasets demonstrate the effectiveness of our approach compared with seven baselines.

The novel coronavirus disease (COVID-19) has crushed daily routines and is still rampaging through the world. Existing solution for nonpharmaceutical interventions usually needs to timely and precisely select a subset of residential urban areas for containment or even quarantine, where the spatial distribution of confirmed cases has been considered as a key criterion for the subset selection. While such containment measure has successfully stopped or slowed down the spread of COVID-19 in some countries, it is criticized for being inefficient or ineffective, as the statistics of confirmed cases are usually time-delayed and coarse-grained. To tackle the issues, we propose C-Watcher, a novel data-driven framework that aims at screening every neighborhood in a target city and predicting infection risks, prior to the spread of COVID-19 from epicenters to the city. In terms of design, C-Watcher collects large-scale long-term human mobility data from Baidu Maps, then characterizes every residential neighborhood in the city using a set of features based on urban mobility patterns. Furthermore, to transfer the firsthand knowledge (witted in epicenters) to the target city before local outbreaks, we adopt a novel adversarial encoder framework to learn “city-invariant” representations from the mobility-related features for precise early detection of high-risk neighborhoods, even before any confirmed cases known, in the target city. We carried out extensive experiments on C-Watcher using the real-data records in the early stage of COVID-19 outbreaks, where the results demonstrate the efficiency and effectiveness of C-Watcher for early detection of high-risk neighborhoods from a large number of cities.

Out-of-town recommendation is designed for those users who leave their home-town areas and visit the areas they have never been to before. It is challenging to recommend Point-of-Interests (POIs) for out-of-town users since the out-of-town check-in behavior is determined by not only the user’s home-town preference but also the user’s travel intention. Besides, the user’s travel intentions are complex and dynamic, which leads to big difficulties in understanding such intentions precisely. In this paper, we propose a TRAvel-INtention-aware Out-of-town Recommendation framework, named TRAINOR. The proposed TRAINOR framework distinguishes itself from existing out-of-town recommenders in three aspects. First, graph neural networks are explored to represent users’ home-town check-in preference and geographical constraints in out-of-town check-in behaviors. Second, a user-specific travel intention is formulated as an aggregation combining home-town preference and generic travel intention together, where the generic travel intention is regarded as a mixture of inherent intentions that can be learned by Neural Topic Model (NTM). Third, a non-linear mapping function, as well as a matrix factorization method, are employed to transfer users’ home-town preference and estimate out-of-town POI’s representation, respectively. Extensive experiments on real-world data sets validate the effectiveness of the TRAINOR framework. Moreover, the learned travel intention can deliver meaningful explanations for understanding a user’s travel purposes.

Deep learning models have demonstrated superior performance in various healthcare applications. However, the major limitation of these deep models is usually the lack of high-quality training data due to the private and sensitive nature of this field. In this study, we propose a novel textual data augmentation method to generate artificial clinical notes in patients’ Electronic Health Records (EHRs) that can be used as additional training data for patient outcomes prediction. Essentially, we fine-tune the generative language model GPT-2 to synthesize labeled text with the original training data. More specifically, We propose a teacher-student framework where we first pre-train a teacher model on the 0riginal data, and then train a student model on the GPT-augmented data under the guidance of the teacher. We evaluate our method on the most common patient outcome, i.e., the 30-day readmission rate. The experimental results show that deep models can improve their predictive performance with the augmented data, indicating the effectiveness of the proposed architecture.

While online advertising is one of the major sources of income for search engines, pumping up the incomes from business advertisements while ensuring the user experience becomes a challenging but emerging area. Designing high-quality advertisements with persuasive content has been proved as a way to increase revenues through improving the Click-Through Rate (CTR). However, it is difficult to scale up the design of high-quality ads, due to the lack of automation in creativity. In this paper, we present Commonsense-Enriched Advertisement on Search Engine (CHASE) --- a system for the automatic generation of persuasive ads. CHASE adopts a specially designed language model that fuses the keywords, commonsense-related texts, and marketing contents to generate persuasive advertisements. Specifically, the language model has been pre-trained using massive contents of explicit knowledge and fine-tuned with well-constructed quasi-parallel corpora with effective control of the proportion of commonsense in the generated ads and fitness to the ads' keywords. The effectiveness of the proposed method CHASE has been verified by real-world web traffics for search and manual evaluation. In A/B tests, the advertisements generated by CHASE would bring 11.13% CTR improvement. The proposed model has been deployed to cover three advertisement domains (which are kid education, psychological counseling, and beauty e-commerce) at Baidu, the world's largest Chinese search engine, with adding revenue of about 1 million RMB (Chinese Yuan) per day.

While recent studies on semi-supervised learning have shown remarkable progress in leveraging both labeled and unlabeled data, most of them presume a basic setting of the model is randomly initialized. In this work, we consider semi-supervised learning and transfer learning jointly, leading to a more practical and competitive paradigm that can utilize both powerful pre-trained models from source domain as well as labeled/unlabeled data in the target domain. To better exploit the value of both pre-trained weights and unlabeled target examples, we introduce adaptive consistency regularization that consists of two complementary components: Adaptive Knowledge Consistency (AKC) on the examples between the source and target model, and Adaptive Representation Consistency (ARC) on the target model between labeled and unlabeled examples. Examples involved in the consistency regularization are adaptively selected according to their potential contributions to the target task. We conduct extensive experiments on popular benchmarks including CIFAR-10, CUB-200, and MURA, by fine-tuning the ImageNet pre-trained ResNet-50 model. Results show that our proposed adaptive consistency regularization outperforms state-of-the-art semi-supervised learning techniques such as Pseudo Label, Mean Teacher, and FixMatch. Moreover, our algorithm is orthogonal to existing methods and thus able to gain additional improvements on top of MixMatch and FixMatch. Our code is available at https://github.com/Walleclipse/Semi-Supervised-Transfer-Learning-Paddle.

Universal style transfer retains styles from reference images in content images. While existing methods have achieved state-of-the-art style transfer performance, they are not aware of the content leak phenomenon that the image content may corrupt after several rounds of stylization process. In this paper, we propose ArtFlow to prevent content leak during universal style transfer. ArtFlow consists of reversible neural flows and an unbiased feature transfer module. It supports both forward and backward inferences and operates in a projection-transfer-reversion scheme. The forward inference projects input images into deep features, while the backward inference remaps deep features back to input images in a lossless and unbiased way. Extensive experiments demonstrate that ArtFlow achieves comparable performance to state-of-the-art style transfer methods while avoiding content leak.

Social networks face a major challenge in the form of rumors and fake news, due to their intrinsic nature of connecting users to millions of others, and of giving any individual the power to post anything.Given the rapid, widespread dissemination of information in social networks, manually detecting suspicious news is sub-optimal.Thus, research on automatic rumor detection has become a necessity.Previous works in the domain have utilized the reply relations between posts, as well as the semantic similarity between the main post and its context, consisting of replies, in order to obtain state-of-the-art performance.In this work, we demonstrate that semantic oppositeness can improve the performance on the task of rumor detection.We show that semantic oppositeness captures elements of discord, which are not properly covered by previous efforts, which only utilize semantic similarity or reply structure.Our proposed model learns both explicit and implicit relations between the main tweet and its replies, by utilizing both semantic similarity and semantic oppositeness.Both of these employ the self-attention mechanism in neural text modeling, with semantic oppositeness utilizing word-level self-attention, and with semantic similarity utilizing post-level self-attention.We show, with extensive experiments on recent data sets for this problem, that our proposed model achieves state-of-theart performance.Further, we show that our model is more resistant to the variances in performance introduced by randomness.

Short text classification is a fundamental task in natural language processing.It is hard due to the lack of context information and labeled data in practice.In this paper, we propose a new method called SHINE, which is based on graph neural network (GNN), for short text classification.First, we model the short text dataset as a hierarchical heterogeneous graph consisting of word-level component graphs which introduce more semantic and syntactic information.Then, we dynamically learn a short document graph that facilitates effective label propagation among similar short texts.Thus, comparing with existing GNN-based methods, SHINE can better exploit interactions between nodes of the same types and capture similarities between short texts.Extensive experiments on various benchmark short text datasets show that SHINE consistently outperforms state-of-the-art methods, especially with fewer labels. 1

Recent literatures have shown that knowledge graph (KG) learning models are highly vulnerable to adversarial attacks. However, there is still a paucity of vulnerability analyses of cross-lingual entity alignment under adversarial attacks. This paper proposes an adversarial attack model with two novel attack techniques to perturb the KG structure and degrade the quality of deep cross-lingual entity alignment. First, an entity density maximization method is employed to hide the attacked entities in dense regions in two KGs, such that the derived perturbations are unnoticeable. Second, an attack signal amplification method is developed to reduce the gradient vanishing issues in the process of adversarial attacks for further improving the attack effectiveness.

While deep learning succeeds in a wide range of tasks, it highly depends on the massive collection of annotated data which is expensive and time-consuming. To lower the cost of data annotation, active learning has been proposed to interactively query an oracle to annotate a small proportion of informative samples in an unlabeled dataset. Inspired by the fact that the samples with higher loss are usually more informative to the model than the samples with lower loss, in this paper we present a novel deep active learning approach that queries the oracle for data annotation when the unlabeled sample is believed to incorporate high loss. The core of our approach is a measurement Temporal Output Discrepancy (TOD) that estimates the sample loss by evaluating the discrepancy of outputs given by models at different optimization steps. Our theoretical investigation shows that TOD lower-bounds the accumulated sample loss thus it can be used to select informative unlabeled samples. On basis of TOD, we further develop an effective unlabeled data sampling strategy as well as an unsupervised learning criterion that enhances model performance by incorporating the unlabeled data. Due to the simplicity of TOD, our active learning approach is efficient, flexible, and task-agnostic. Extensive experimental results demonstrate that our approach achieves superior performances than the state-of-the-art active learning methods on image classification and semantic segmentation tasks.

A recent study has shown that graph matching models are vulnerable to adversarial manipulation of their input which is intended to cause a mismatching. Nevertheless, there is still a lack of a comprehensive solution for further enhancing the robustness of graph matching against adversarial attacks. In this paper, we identify and study two types of unique topology attacks in graph matching: inter-graph dispersion and intra-graph assembly attacks. We propose an integrated defense model, IDRGM, for resilient graph matching with two novel defense techniques to defend against the above two attacks simultaneously. A detection technique of inscribed simplexes in the hyperspheres consisting of multiple matched nodes is proposed to tackle inter-graph dispersion attacks, in which the distances among the matched nodes in multiple graphs are maximized to form regular simplexes. A node separation method based on phase-type distribution and maximum likelihood estimation is developed to estimate the distribution of perturbed graphs and separate the nodes within the same graphs over a wide space, for defending intra-graph assembly attacks, such that the interference from the similar neighbors of the perturbed nodes is significantly reduced. We evaluate the robustness of our IDRGM model on real datasets against state-of-the-art algorithms.

Recent findings have shown multiple graph learning models, such as graph classification and graph matching, are highly vulnerable to adversarial attacks, i.e. small input perturbations in graph structures and node attributes can cause the model failures. Existing defense techniques often defend specific attacks on particular multiple graph learning tasks. This paper proposes an attack-agnostic graph-adaptive 1-Lipschitz neural network, ERNN, for improving the robustness of deep multiple graph learning while achieving remarkable expressive power. A K_l-Lipschitz Weibull activation function is designed to enforce the gradient norm as K_l at layer l. The nearest matrix orthogonalization and polar decomposition techniques are utilized to constraint the weight norm as 1/K_l and make the norm-constrained weight close to the original weight. The theoretical analysis is conducted to derive lower and upper bounds of feasible K_l under the 1-Lipschitz constraint. The combination of norm-constrained weight and activation function leads to the 1-Lipschitz neural network for expressive and robust multiple graph learning.

Real estate appraisal refers to the process of developing an unbiased opinion for real property's market value, which plays a vital role in decision-making for various players in the marketplace (e.g., real estate agents, appraisers, lenders, and buyers). However, it is a non-trivial task for accurate real estate appraisal because of three major challenges: (1) The complicated influencing factors for property value; (2) The asynchronously spatiotemporal dependencies among real estate transactions; (3) The diversified correlations between residential communities. To this end, we propose a Multi-Task Hierarchical Graph Representation Learning (MugRep) framework for accurate real estate appraisal. Specifically, by acquiring and integrating multi-source urban data, we first construct a rich feature set to profile the real estate from multiple perspectives~(e.g., geographical distribution, human mobility distribution, and resident demographics distribution). Then, an evolving real estate transaction graph and a corresponding event graph convolution module are proposed to incorporate asynchronously spatiotemporal dependencies among real estate transactions. Moreover, to further incorporate valuable knowledge from the view of residential communities, we devise a hierarchical heterogeneous community graph convolution module to capture diversified correlations between residential communities. Finally, an urban district partitioned multi-task learning module is introduced to generate differently distributed value opinions for real estate. Extensive experiments on two real-world datasets demonstrate the effectiveness of MugRep and its components and features.

Drug discovery often relies on the successful prediction of protein-ligand binding affinity. Recent advances have shown great promise in applying graph neural networks (GNNs) for better affinity prediction by learning the representations of protein-ligand complexes. However, existing solutions usually treat protein-ligand complexes as topological graph data, thus the biomolecular structural information is not fully utilized. The essential long-range interactions among atoms are also neglected in GNN models. To this end, we propose a structure-aware interactive graph neural network (SIGN) which consists of two components: polar-inspired graph attention layers (PGAL) and pairwise interactive pooling (PiPool). Specifically, PGAL iteratively performs the node-edge aggregation process to update embeddings of nodes and edges while preserving the distance and angle information among atoms. Then, PiPool is adopted to gather interactive edges with a subsequent reconstruction loss to reflect the global interactions. Exhaustive experimental study on two benchmarks verifies the superiority of SIGN.

In modern internet industries, deep learning based recommender systems have became an indispensable building block for a wide spectrum of applications, such as search engine, news feed, and short video clips. However, it remains challenging to carry the well-trained deep models for online real-time inference serving, with respect to the time-varying web-scale traffics from billions of users, in a cost-effective manner. In this work, we present JIZHI - a Model-as-a-Service system - that per second handles hundreds of millions of online inference requests to huge deep models with more than trillions of sparse parameters, for over twenty real-time recommendation services at Baidu, Inc. In JIZHI, the inference workflow of every recommendation request is transformed to a Staged Event-Driven Pipeline (SEDP), where each node in the pipeline refers to a staged computation or I/O intensive task processor. With traffics of real-time inference requests arrived, each modularized processor can be run in a fully asynchronized way and managed separately. Besides, JIZHI introduces the heterogeneous and hierarchical storage to further accelerate the online inference process by reducing unnecessary computations and potential data access latency induced by ultra-sparse model parameters. Moreover, an intelligent resource manager has been deployed to maximize the throughput of JIZHI over the shared infrastructure by searching the optimal resource allocation plan from historical logs and fine-tuning the load shedding policies over intermediate system feedback. Extensive experiments have been done to demonstrate the advantages of JIZHI from the perspectives of end-to-end service latency, system-wide throughput, and resource consumption. Since launched in July 2019, JIZHI has helped Baidu saved more than ten million US dollars in hardware and utility costs per year while handling 200% more traffics without sacrificing the inference efficiency.

Hang Hua, Xingjian Li, Dejing Dou, Chengzhong Xu, Jiebo Luo. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Label noise and class imbalance are two major issues coexisting in real-world datasets. To alleviate the two issues, state-of-the-art methods reweight each instance by leveraging a small amount of clean and unbiased data. Yet, these methods overlook class-level information within each instance, which can be further utilized to improve performance. To this end, in this paper, we propose Generalized Data Weighting (GDW) to simultaneously mitigate label noise and class imbalance by manipulating gradients at the class level. To be specific, GDW unrolls the loss gradient to class-level gradients by the chain rule and reweights the flow of each gradient separately. In this way, GDW achieves remarkable performance improvement on both issues. Aside from the performance gain, GDW efficiently obtains class-level weights without introducing any extra computational cost compared with instance weighting methods. Specifically, GDW performs a gradient descent step on class-level weights, which only relies on intermediate gradients. Extensive experiments in various settings verify the effectiveness of GDW. For example, GDW outperforms state-of-the-art methods by $2.56\%$ under the $60\%$ uniform noise setting in CIFAR10. Our code is available at https://github.com/GGchen1997/GDW-NIPS2021.

Molecular property prediction plays a fundamental role in drug discovery to identify candidate molecules with target properties. However, molecular property prediction is essentially a few-shot problem, which makes it hard to use regular machine learning models. In this paper, we propose Property-Aware Relation networks (PAR) to handle this problem. In comparison to existing works, we leverage the fact that both relevant substructures and relationships among molecules change across different molecular properties. We first introduce a property-aware embedding function to transform the generic molecular embeddings to substructure-aware space relevant to the target property. Further, we design an adaptive relation graph learning module to jointly estimate molecular relation graph and refine molecular embeddings w.r.t. the target property, such that the limited labels can be effectively propagated among similar molecules. We adopt a meta-learning strategy where the parameters are selectively updated within tasks in order to model generic and property-aware knowledge separately. Extensive experiments on benchmark molecular property prediction datasets show that PAR consistently outperforms existing methods and can obtain property-aware molecular embeddings and model molecular relation graph properly.

Despite achieving remarkable efficiency, traditional network pruning techniques often follow manually-crafted heuristics to generate pruned sparse networks. Such heuristic pruning strategies are hard to guarantee that the pruned networks achieve test accuracy comparable to the original dense ones. Recent works have empirically identified and verified the Lottery Ticket Hypothesis (LTH): a randomly-initialized dense neural network contains an extremely sparse subnetwork, which can be trained to achieve similar accuracy to the former. Due to the lack of theoretical evidence, they often need to run multiple rounds of expensive training and pruning over the original large networks to discover the sparse subnetworks with low accuracy loss. By leveraging dynamical systems theory and inertial manifold theory, this work theoretically verifies the validity of the LTH. We explore the possibility of theoretically lossless pruning as well as one-time pruning, compared with existing neural network pruning and LTH techniques. We reformulate the neural network optimization problem as a gradient dynamical system and reduce this high-dimensional system onto inertial manifolds to obtain a low-dimensional system regarding pruned subnetworks. We demonstrate the precondition and existence of pruned subnetworks and prune the original networks in terms of the gap in their spectrum that make the subnetworks have the smallest dimensions.

Feature selection is a fundamental component of data mining, aiming to select optimal feature subsets for downstream task.Recently, an emerging feature selection method called reinforced feature selection applies reinforcement learning into feature selection.Reinforced Feature Selection (RFS) automates feature selection process and can effectively find the optimal subset.Generally, RFS can be categorized into single-agent RFS and multi-agent RFS.Single-agent RFS uses one reinforcement learning agent to select features, but its action space is exponentially-increasing with feature number and can merely obtain local optima.Multi-agent RFS uses multiple agents to select features; this method can achieve global optima, but it needs to optimize as many policy networks as feature number which costs huge computational resources and thus becomes computationally inefficient.This dilemma naturally leads to a research question: How can we synthesize the advantages of single-agent RFS and multi-agent RFS while avoiding their disadvantages?To answer this question, we propose a Group-based Interactive Reinforced Feature Selection (GIRFS) framework.This framework balances single-agent RFS and multi-agent RFS for better feature selection.Specifically, we formulate the feature selection problem into a group-based RFS problem.In this formulation, we first assign the given features into several groups based on feature similarity measurement.Then, we create agents for each group, where each agent decides to select/deselect features in its corresponding group.This design balances the size of action space and number of policy networks and thus makes RFS more effective and efficient.Moreover, to further improve learning efficiency, we propose a hierarchical teacher-like trainer to provide external action advice for agents.This trainer provides advice by intra-group selection and inter-group selection and fuses knowledge from mRMR and decision tree to help agents explore and learn.Finally, we present extensive experiments on real-world datasets to demonstrate the improved performances of our method.

Unplanned intensive care unit (ICU) readmission rate is an important metric for evaluating the quality of hospital care. Efficient and accurate prediction of ICU readmission risk can not only help prevent patients from inappropriate discharge and potential dangers, but also reduce associated costs of healthcare. In this paper, we propose a new method that uses medical text of Electronic Health Records (EHRs) for prediction, which provides an alternative perspective to previous studies that heavily depend on numerical and time-series features of patients. More specifically, we extract discharge summaries of patients from their EHRs, and represent them with multiview graphs enhanced by an external knowledge graph. Graph convolutional networks are then used for representation learning. Experimental results prove the effectiveness of our method, yielding state-of-the-art performance for this task.

Electric Vehicle (EV) has become a preferable choice in the modern transportation system due to its environmental and energy sustainability. However, in many large cities, EV drivers often fail to find the proper spots for charging, because of the limited charging infrastructures and the spatiotemporally unbalanced charging demands. Indeed, the recent emergence of deep reinforcement learning provides great potential to improve the charging experience from various aspects over a long-term horizon. In this paper, we propose a framework, named Multi-Agent Spatio-Temporal Reinforcement Learning (Master), for intelligently recommending public accessible charging stations by jointly considering various long-term spatiotemporal factors. Specifically, by regarding each charging station as an individual agent, we formulate this problem as a multi-objective multi-agent reinforcement learning task. We first develop a multi-agent actor-critic framework with the centralized attentive critic to coordinate the recommendation between geo-distributed agents. Moreover, to quantify the influence of future potential charging competition, we introduce a delayed access strategy to exploit the knowledge of future charging competition during training. After that, to effectively optimize multiple learning objectives, we extend the centralized attentive critic to multi-critics and develop a dynamic gradient re-weighting strategy to adaptively guide the optimization direction. Finally, extensive experiments on two real-world datasets demonstrate that Master achieves the best comprehensive performance compared with nine baseline approaches.

The operation and management of intelligent transportation systems (ITS), such as traffic monitoring, relies on real-time data aggregation of vehicular traffic information, including vehicular types (e.g., cars, trucks, and buses), in the critical roads and highways. While traditional approaches based on vehicular-embedded GPS sensors or camera networks would either invade drivers’ privacy or require high deployment cost, this article introduces a low-cost method, namely,SenseMag, to recognize the vehicular type using a pair of noninvasive magnetic sensors deployed on the straight road section.SenseMagfilters out noises and segments received magnetic signals by the exact time points that the vehicle arrives or departs from every sensor node. Furthermore,SenseMagadopts a hierarchical recognition model to first estimate the speed/velocity, then identify the length of the vehicle using the predicted speed, sampling cycles, and the distance between the sensor nodes. With the vehicle length identified and the temporal/spectral features extracted from the magnetic signals,SenseMagclassifies the types of vehicles accordingly. Some semiautomated learning techniques have been adopted for the design of filters, features, and the choice of hyperparameters. Extensive experiment based on real-word field deployment (on the highways in Shenzhen, China) shows thatSenseMagsignificantly outperforms the existing methods in both classification accuracy and the granularity of vehicle types (i.e., seven types bySenseMagversus four types by the existing work in comparisons). To be specific, our field experiment results validate thatSenseMagis with at least 90% vehicle type classification accuracy and less than 5% vehicle length classification error.

Ensemble learning is a widely used technique to train deep convolutional neural networks (CNNs) for improved robustness and accuracy. While existing algorithms usually first train multiple diversified networks and then assemble these networks as an aggregated classifier, we propose a novel learning paradigm, namely, “In-Network Ensemble” ( INE ) that incorporates the diversity of multiple models through training a SINGLE deep neural network. Specifically, INE segments the outputs of the CNN into multiple independent classifiers, where each classifier is further fine-tuned with better accuracy through a so-called diversified knowledge distillation process . We then aggregate the fine-tuned independent classifiers using an Averaging-and-Softmax operator to obtain the final ensemble classifier. Note that, in the supervised learning settings, INE starts the CNN training from random, while, under the transfer learning settings, it also could start with a pre-trained model to incorporate the knowledge learned from additional datasets. Extensive experiments have been done using eight large-scale real-world datasets, including CIFAR, ImageNet, and Stanford Cars, among others, as well as common deep network architectures such as VGG, ResNet, and Wide ResNet. We have evaluated the method under two tasks: supervised learning and transfer learning. The results show that INE outperforms the state-of-the-art algorithms for deep ensemble learning with improved accuracy.

Stochastic Gradient Langevin Dynamics (SGLD) have been widely used for Bayesian sampling from certain probability distributions, incorporating derivatives of the log-posterior. With the derivative evaluation of the log-posterior distribution, SGLD methods generate samples from the distribution through performing as a thermostats dynamics that traverses over gradient flows of the log-posterior with certainly controllable perturbation. Even when the density is not known, existing solutions still can first learn the kernel density models from the given datasets, then produce new samples using the SGLD over the kernel density derivatives. In this work, instead of exploring new samples from kernel spaces, a novel SGLD sampler, namely, Randomized Measurement Langevin Dynamics (RMLD) is proposed to sample the high-dimensional sparse representations from the spectral domain of a given dataset. Specifically, given a random measurement matrix for sparse coding, RMLD first derives a novel likelihood evaluator of the probability distribution from the loss function of LASSO, then samples from the high-dimensional distribution using stochastic Langevin dynamics with derivatives of the logarithm likelihood and Metropolis–Hastings sampling. In addition, new samples in low-dimensional measuring spaces can be regenerated using the sampled high-dimensional vectors and the measurement matrix. The algorithm analysis shows that RMLD indeed projects a given dataset into a high-dimensional Gaussian distribution with Laplacian prior, then draw new sparse representation from the dataset through performing SGLD over the distribution. Extensive experiments have been conducted to evaluate the proposed algorithm using real-world datasets. The performance comparisons on three real-world applications demonstrate the superior performance of RMLD beyond baseline methods.

In this paper, we extend the classic MaxOut strategy, originally designed for Multiple Layer Preceptors (MLPs), intoCOnvolutionalMaxOut (COMO) — a new strategy making deep convolutional neural networks wider with parameter efficiency. Compared to the existing solutions, such as ResNeXt for ResNet or Inception for VGG-alikes, COMO works well on both linear architectures and the ones with skipped connections and residual blocks. More specifically, COMO adopts a novelsplit-transform-mergeparadigm that extends the layers withspatial resolution reductioninto multiple parallel splits. For the layer with COMO, each split passes the input feature maps through a4D convolution operatorwith independentbatch normalization operatorsfor transformation, then merge into the aggregated output of the original sizes throughmax-pooling. Such a strategy is expected to tackle the potential classification accuracy degradation due to the spatial resolution reduction, by incorporating the multiple splits and max-pooling-based feature selection. Our experiment using a wide range of deep architectures shows that COMO can significantly improve the classification accuracy of ResNet/VGG-alike networks based on a large number of benchmark datasets. COMO further outperforms the existing solutions, e.g., Inceptions, ResNeXts, SE-ResNet, and Xception, that make networks wider, and it dominates in the comparison of accuracy versus parameter sizes.

Definition Extraction (DE) is one of the well-known topics in Information Extraction that aims to identify terms and their corresponding definitions in unstructured texts. This task can be formalized either as a sentence classification task (i.e., containing term-definition pairs or not) or a sequential labeling task (i.e., identifying the boundaries of the terms and definitions). The previous works for DE have only focused on one of the two approaches, failing to model the inter-dependencies between the two tasks. In this work, we propose a novel model for DE that simultaneously performs the two tasks in a single framework to benefit from their inter-dependencies. Our model features deep learning architectures to exploit the global structures of the input sentences as well as the semantic consistencies between the terms and the definitions, thereby improving the quality of the representation vectors for DE. Besides the joint inference between sentence classification and sequential labeling, the proposed model is fundamentally different from the prior work for DE in that the prior work has only employed the local structures of the input sentences (i.e., word-to-word relations), and not yet considered the semantic consistencies between terms and definitions. In order to implement these novel ideas, our model presents a multi-task learning framework that employs graph convolutional neural networks and predicts the dependency paths between the terms and the definitions. We also seek to enforce the consistency between the representations of the terms and definitions both globally (i.e., increasing semantic consistency between the representations of the entire sentences and the terms/definitions) and locally (i.e., promoting the similarity between the representations of the terms and the definitions). The extensive experiments on three benchmark datasets demonstrate the effectiveness of our approach.1

Relation Extraction (RE) is one of the fundamental tasks in Information Extraction. The goal of this task is to find the semantic relations between entity mentions in text. It has been shown in many previous work that the structure of the sentences (i.e., dependency trees) can provide important information/features for the RE models. However, the common limitation of the previous work on RE is the reliance on some external parsers to obtain the syntactic trees for the sentence structures. On the one hand, it is not guaranteed that the independent external parsers can offer the optimal sentence structures for RE and the customized structures for RE might help to further improve the performance. On the other hand, the quality of the external parsers might suffer when applied to different domains, thus also affecting the performance of the RE models on such domains. In order to overcome this issue, we introduce a novel method for RE that simultaneously induces the structures and predicts the relations for the input sentences, thus avoiding the external parsers and potentially leading to better sentence structures for RE. Our general strategy to learn the RE-specific structures is to apply two different methods to infer the structures for the input sentences (i.e., two views). We then introduce several mechanisms to encourage the structure and semantic consistencies between these two views so the effective structure and semantic representations for RE can emerge. We perform extensive experiments on the ACE 2005 and SemEval 2010 datasets to demonstrate the advantages of the proposed method, leading to the state-of-the-art performance on such datasets.

External knowledge is often useful for natural language understanding tasks. We introduce a contextual text representation model called Conceptual-Contextual (CC) embeddings, which incorporates structured knowledge into text representations. Unlike entity embedding methods, our approach encodes a knowledge graph into a context model. CC embeddings can be easily reused for a wide range of tasks in a similar fashion to pre-trained language models. Our model effectively encodes the huge UMLS database by leveraging semantic generalizability. Experiments on electronic health records (EHRs) and medical text processing benchmarks showed our model gives a major boost to the performance of supervised medical NLP tasks.

This paper studies the task of Relation Extraction (RE) that aims to identify the semantic relations between two entity mentions in text.In the deep learning models for RE, it has been beneficial to incorporate the syntactic structures from the dependency trees of the input sentences.In such models, the dependency trees are often used to directly structure the network architectures or to obtain the dependency relations between the word pairs to inject the syntactic information into the models via multi-task learning.The major problems with these approaches are the lack of generalization beyond the syntactic structures in the training data or the failure to capture the syntactic importance of the words for RE.In order to overcome these issues, we propose a novel deep learning model for RE that uses the dependency trees to extract the syntax-based importance scores for the words, serving as a tree representation to introduce syntactic information into the models with greater generalization.In particular, we leverage Ordered-Neuron Long-Short Term Memory Networks (ON-LSTM) to infer the model-based importance scores for RE for every word in the sentences that are then regulated to be consistent with the syntax-based scores to enable syntactic information injection.We perform extensive experiments to demonstrate the effectiveness of the proposed method, leading to the state-of-the-art performance on three RE benchmark datasets.

As it is difficult to directly share data among different organizations, data federation brings new opportunities to the data-related cooperation among different organizations by providing abstract data interfaces. With the development of Cloud computing, organizations store data on the Cloud to achieve elasticity and scalability for data processing. The existing data placement approaches generally only consider one aspect, which is either communication cost or time cost, and do not consider the features of jobs that process the data. In this paper, we propose an approach to enable secure data processing on the Cloud with the data from different organizations. The approach consists of a data federation platform for secure data processing on the Cloud named FedCube and a greedy data placement algorithm that creates a plan to store data on the Cloud in order to achieve multiple objectives based on a cost model. The cost model is composed of two objectives, i.e., reducing both monetary cost and execution time. We present an experimental evaluation by comparing our data placement algorithm with the existing methods based on the data federation platform. The experiments show that our proposed algorithm significantly reduce the total cost (up to 69.8%).

As the recent COVID-19 outbreak rapidly expands all over the world, various containment measures have been carried out to fight against the COVID-19 pandemic. In Mainland China, the containment measures consist of three types, i.e., Wuhan travel ban, intra-city quarantine and isolation, and intercity travel restriction. In order to carry out the measures, local economy and information acquisition play an important role. In this paper, we investigate the correlation of local economy and the information acquisition on the execution of containment measures to fight against the COVID-19 pandemic in Mainland China. First, we use a parsimonious model, i.e., SIR-X model to estimate the parameters, which represent the execution of intra-city quarantine and isolation in major cities of Mainland China. In order to understand the execution of intra-city quarantine and isolation, we analyze the correlation between the representative parameters including local economy, mobility, and information acquisition. To this end, we collect the data of Gross Domestic Product (GDP), the inflows from Wuhan and outflows, and the COVID-19 related search frequency from a widely-used Web mapping service, i.e., Baidu Maps, and Web search engine, i.e., Baidu Search Engine, in Mainland China. Based on the analysis, we confirm the strong correlation between the local economy and the execution of information acquisition in major cities of Mainland China. We further evidence that, although the cities with high GDP per capita attract more inflows from Wuhan, people are more likely to conduct the quarantine measure and to reduce travelling to other cities. Finally, the correlation analysis using search data shows that well-informed individuals are likely to carry out containment measures.

Network alignment techniques, which aim to identify the same entities across multiple networks, often suffer challenges from feature inconsistency to transitivity law preservation. This paper presents a purely unsupervised network alignment method, KEMINA, with three original contributions. First, in order to address the feature inconsistency issue, an adversarial kernel embedding technique is proposed to extract network-invariant information among multiple networks without prior alignment knowledge, and project them into the common embedding space. Second, a multinomial generative adversarial network (GAN) model is developed to train multiple network alignment tasks simultaneously in an unsupervised manner with preserving the transitivity law property. Third but last, a variational inference model is designed to alleviate the data sparsity and inadequate training issues by filling realistic detail for vertices with sparse features and generating real-looking supplementary vertex samples within limited training opportunity of each pair of source and target networks.

Network representation learning (NRL) is crucial in the area of graph learning.Recently, graph autoencoders and its variants have gained much attention and popularity among various types of node embedding approaches.Most existing graph autoencoder-based methods aim to minimize the reconstruction errors of the input network while not explicitly considering the semantic relatedness between nodes.In this paper, we propose a novel network embedding method which models the consistency across different views of networks.More specifically, we create a second view from the input network which captures the relation between nodes based on node content and enforce the latent representations from the two views to be consistent by incorporating a multiview adversarial regularization module.The experimental studies on benchmark datasets prove the effectiveness of this method, and demonstrate that our method compares favorably with the state-of-the-art algorithms on challenging tasks such as link prediction and node clustering.We also evaluate our method on a real-world application, i.e., 30-day unplanned ICU readmission prediction, and achieve promising results compared with several baseline methods.

Targeted opinion word extraction (TOWE) is a sub-task of aspect based sentiment analysis (ABSA) which aims to find the opinion words for a given aspect-term in a sentence.Despite their success for TOWE, the current deep learning models fail to exploit the syntactic information of the sentences that have been proved to be useful for TOWE in the prior research.In this work, we propose to incorporate the syntactic structures of the sentences into the deep learning models for TOWE, leveraging the syntax-based opinion possibility scores and the syntactic connections between the words.We also introduce a novel regularization technique to improve the performance of the deep learning models based on the representation distinctions between the words in TOWE.The proposed model is extensively analyzed and achieves the state-of-the-art performance on four benchmark datasets.

Recent studies have shown that graph mining models are vulnerable to adversarial attacks. This paper proposes a robust meta network embedding framework, RoMNE, which improves the robustness of multiple network embedding on adversarial noisy networks while preserving the utility on original clean ones. First, we propose a generic meta learning based multiple network embedding model that can quickly adapt it to new embedding tasks on a variety of network data with only a small number of parameter and training updates. Second, Gumbel estimator and Gaussian smoothing techniques are introduced to implement differentiable approximation for optimizing non-differential objective of effective adversarial attacks. Last but not least, the adversarial attack and defense models are integrated into a dynamic adversarial training model. The competition of two models helps the latter be robust to adversarial attacks.

Local community detection for a given set of query nodes attracts much research attention recently. The query nodes play essential roles in the detection effectiveness. Existing methods perform well when a query node is from the target community core region. However, they struggle with the query-bias issue and especially perform unsatisfactorily when the query nodes come from different communities or when certain query nodes are from communities overlapping region or community boundary region. To address above issues, we consider from a new angle, to replace these original “intractable” query nodes with new detection-friendly query nodes. In this paper, we propose an effective ATP (Amplified Topology Potential) algorithm to detect core nodes of the target communities w.r.t. original query nodes. For one query node, ATP first builds a query-oriented topology potential field around the query node by aggregating random walk with restart scores. Then it amplifies the topology potential value to make core nodes of target communities easily distinguished. Graph-size-independent fast approximation strategies are also proposed together with sound theoretical foundations. Extensive experiments on four real networks using ten state-of-the-art local community detection methods verify the improvement in detection effectiveness and efficiency by the replacing strategy for the tough query cases.

Fine-tuning the deep convolution neural network (CNN) using a pre-trained model helps transfer knowledge learned from larger datasets to the target task. While the accuracy could be largely improved even when the training dataset is small, the transfer learning outcome is similar with the pre-trained one with closed CNN weights[17], as the backpropagation here brings less updates to deeper CNN layers. In this work, we propose RIFLE - a simple yet effective strategy that deepens backpropagation in transfer learning settings, through periodically ReInitializing the Fully-connected LayEr with random scratch during the fine-tuning procedure. RIFLE brings significant perturbation to the backpropagation process and leads to deep CNN weights update, while the affects of perturbation can be easily converged throughout the overall learning procedure. The experiments show that the use of RIFLE significantly improves deep transfer learning accuracy on a wide range of datasets, outperforming known tricks for the similar purpose, such as dropout, dropconnect, stochastic depth, and cyclic learning rate, under the same settings with 0.5%-2% higher testing accuracy. Empirical cases and ablation studies further indicate RIFLE brings meaningful updates to deep CNN layers with accuracy improved.

In this paper, we aim to develop a scalable algorithm to preserve differential privacy (DP) in adversarial learning for deep neural networks (DNNs), with certified robustness to adversarial examples. By leveraging the sequential composition theory in DP, we randomize both input and latent spaces to strengthen our certified robustness bounds. To address the trade-off among model utility, privacy loss, and robustness, we design an original adversarial objective function, based on the post-processing property in DP, to tighten the sensitivity of our model. A new stochastic batch training is proposed to apply our mechanism on large DNNs and datasets, by bypassing the vanilla iterative batch-by-batch training in DP DNNs. An end-to-end theoretical analysis and evaluations show that our mechanism notably improves the robustness and scalability of DP DNNs.

Generation of high-quality person images is challenging, due to the sophisticated entanglements among image factors, e.g., appearance, pose, foreground, background, local details, global structures, etc. In this paper, we present a novel end-to-end framework to generate realistic person images based on given person poses and appearances. The core of our framework is a novel generator called Appearance-aware Pose Stylizer (APS) which generates human images by coupling the target pose with the conditioned person appearance progressively. The framework is highly flexible and controllable by effectively decoupling various complex person image factors in the encoding phase, followed by re-coupling them in the decoding phase. In addition, we present a new normalization method named adaptive patch normalization, which enables region-specific normalization and shows a good performance when adopted in person image generation model. Experiments on two benchmark datasets show that our method is capable of generating visually appealing and realistic-looking results using arbitrary image and pose inputs.

In this paper, we introduce a novel interpreting framework that learns an interpretable model based on an ontology-based sampling technique to explain agnostic prediction models. Different from existing approaches, our algorithm considers contextual correlation among words, described in domain knowledge ontologies, to generate semantic explanations. To narrow down the search space for explanations, which is a major problem of long and complicated text data, we design a learnable anchor algorithm, to better extract explanations locally. A set of regulations is further introduced, regarding combining learned interpretable representations with anchors to generate comprehensible semantic explanations. An extensive experiment conducted on two real-world datasets shows that our approach generates more precise and insightful explanations compared with baseline approaches.

Discriminatively localizing sounding objects in cocktail-party, i.e., mixed sound scenes, is commonplace for humans, but still challenging for machines. In this paper, we propose a two-stage learning framework to perform self-supervised class-aware sounding object localization. First, we propose to learn robust object representations by aggregating the candidate sound localization results in the single source scenes. Then, class-aware object localization maps are generated in the cocktail-party scenarios by referring the pre-learned object knowledge, and the sounding objects are accordingly selected by matching audio and visual object category distributions, where the audiovisual consistency is viewed as the self-supervised signal. Experimental results in both realistic and synthesized cocktail-party videos demonstrate that our model is superior in filtering out silent objects and pointing out the location of sounding objects of different classes. Code is available at https://github.com/DTaoo/Discriminative-Sounding-Objects-Localization.

Despite achieving remarkable performance, deep graph learning models, such as node classification and network embedding, suffer from harassment caused by small adversarial perturbations. However, the vulnerability analysis of graph matching under adversarial attacks has not been fully investigated yet. This paper proposes an adversarial attack model with two novel attack techniques to perturb the graph structure and degrade the quality of deep graph matching: (1) a kernel density estimation approach is utilized to estimate and maximize node densities to derive imperceptible perturbations, by pushing attacked nodes to dense regions in two graphs, such that they are indistinguishable from many neighbors; and (2) a meta learning-based projected gradient descent method is developed to well choose attack starting points and to improve the search performance for producing effective perturbations. We evaluate the effectiveness of the attack model on real datasets and validate that the attacks can be transferable to other graph learning models.

Event factuality prediction (EFP) is the task of assessing the degree to which an event mentioned in a sentence has happened.For this task, both syntactic and semantic information are crucial to identify the important context words.The previous work for EFP has only combined these information in a simple way that cannot fully exploit their coordination.In this work, we introduce a novel graph-based neural network for EFP that can integrate the semantic and syntactic information more effectively.Our experiments demonstrate the advantage of the proposed model for EFP.

Unsupervised word embeddings have become a popular approach of word representation in NLP tasks.However there are limitations to the semantics represented by unsupervised embeddings, and inadequate fine-tuning of embeddings can lead to suboptimal performance.We propose a novel learning technique called Delta Embedding Learning, which can be applied to general NLP tasks to improve performance by optimized tuning of the word embeddings.A structured regularization is applied to the embeddings to ensure they are tuned in an incremental way.As a result, the tuned word embeddings become better word representations by absorbing semantic information from supervision without "forgetting."We apply the method to various NLP tasks and see a consistent improvement in performance.Evaluation also confirms the tuned word embeddings have better semantic properties.

Fake news and rumors constitute a major problem in social networks recently. Due to the fast information propagation in social networks, it is inefficient to use human labor to detect suspicious news. Automatic rumor detection is thus necessary to prevent devastating effects of rumors on the individuals and society. Previous work has shown that in addition to the content of the news/posts and their contexts (i.e., replies), the relations or connections among those components are important to boost the rumor detection performance. In order to induce such relations between posts and contexts, the prior work has mainly relied on the inherent structures of the social networks (e.g., direct replies), ignoring the potential semantic connections between those objects. In this work, we demonstrate that such semantic relations are also helpful as they can reveal the implicit structures to better capture the patterns in the contexts for rumor detection. We propose to employ the self-attention mechanism in neural text modeling to achieve the semantic structure induction for this problem. In addition, we introduce a novel method to preserve the important information of the main news/posts in the final representations of the entire threads to further improve the performance for rumor detection. Our method matches the main post representations and the thread representations by ensuring that they predict the same latent labels in a multitask learning framework. The extensive experiments demonstrate the effectiveness of the proposed model for rumor detection, yielding the state-of-the-art performance on recent datasets for this problem.

Network embedding techniques aim to encode each vertex/edge as a low-dimensional vector, enabling easy integration with existing graph mining algorithms. This paper presents a novel network embedding framework, VEEMBEDCLASS, that combines local vertex/edge embedding with deep matrix fusion and Siamese multi-label classification for facilitating classification-based local network embedding. First, we propose to perform the embeddings of each vertex/edge on K local vertex/edge embedding models respectively, with the joint optimization by considering both intra-class and inter-class correlations, to learn their latent local features on each class. The deep matrix fusion technique is developed to preserve the first-order and second-order proximity of vertices and edges on each of K classes simultaneously. Second, a Student t-distribution based Siamese multi-label classification method is designed to train associated vertices and edges with similar local characteristics together and learn their class membership probabilities, in response to the power-law vertex degree distribution widespread in real graphs. A principle of vertex-edge homophily is introduced to guarantee that the common edge/vertex shared by two associated vertices/edges and themselves are similar in terms of both structural correlations and class memberships. Finally, we integrate local vertex/edge embedding and Siamese multi-label classification into a unified model by mutually enhancing each other.

Real-world graphs are usually very sparse in terms of inadequate edges and labels as well as have poor quality due to a large amount of noisy data. In this paper, we propose a classification-based local vertex ranking architecture through dual generative adversarial networks in the semi-supervised setting, DQGAN, for analyzing sparse noisy graphs with rarely labeled data. First, we develop a quadruple generative adversarial ClassNet model to address the noisy data and data sparsity issues as well as to classify each vertex into K classes by automatically creating imaginary/real-looking supplementary labeled vertices with the quite different/similar distributions as real vertices, without the high cost of multi-step graph propagation, heterogeneous graph mining, and iterative weight learning. In addition, the vertex label vicinity is incorporated into the classification model to capture the pairwise vertex closeness based on the labeling and align the vertex label vicinity with the well-known vertex homophily for preserving the original structural semantics in the classification space. Second, we present a quintuple generative adversarial RankNet framework to locally rank each vertex on each of K classes by designing the game of multiple competitors utilizing the mix of real and noisy data to fight against each other, for improving the robustness of local vertex ranking to noisy data with few help from human efforts. The cycle ranking consistency strategy is designed to make the ranking quality verifiable through the bidirectional information-lossless translations between the original features and the ranking features. We propose to utilize the relaxed local PageRank property to produce high-quality local vertex ranking results in the context of information networks. Third but last, extensive evaluation on real graph datasets demonstrates that DQGAN outperforms existing representative methods in terms of both classification and ranking in the semi-supervised setting.

In this paper, we present a community-focused drug abuse monitoring and supporting system, called DrugTracker, that utilizes social media and geospatial data in near real-time. Through the system, users can: (1) Detect drug abuse risk behaviors from social media platforms, e.g., Twitter; (2) Analyze drug abuse risk behaviors by querying consolidated and live datasets with keywords, spatial entities, and time constraints; and (3) Explore the query results and associated data through a web-based user interface in thematic choropleth, heatmap, and statistical charts. To protect the privacy of the Twitter users, whose data is collected, the system automatically hides the re-identification elements in tweets and aggregates the geo-tags into areas such as census tracts. For the demonstration purpose, our DrugTracker system is populated with a database that contains about 10 million tweets from the year 2017, that were annotated as drug abuse risk behavior positive by our deep learning model.

Network alignment, which aims to learn a matching between the same entities across multiple information networks, often suffers challenges from feature inconsistency, high-dimensional features, to unstable alignment results. This paper presents a novel network alignment framework, RANA, that combines dual generative adversarial network (GAN) techniques to match the distributions of two networks based on two dimensions of distance and shape. First, we propose an adversarial network distribution matching model to perform the bidirectional cross-network alignment translations between two networks, such that the cross-network transformed distributions of two networks move closer to each other and finally meet with each other halfway. In addition, a homophily consistency loss is introduced to maintain the vertex homophily consistency between pairwise vertices on two networks in both the embedding space. Second, in order to address the feature inconsistency issue, we integrate a dual adversarial autoencoder module with an adversarial two-class classification model together to twist the cross-network transformed distributions of two networks, such that two distributions could have the same shape. This facilitates the translations of the distributions of two networks in the adversarial network distribution matching model. Moreover, a semantic preservation loss is introduced to preserve the original embedding semantics of one network when this network is translated to another network and returned to itself. Third but last, the competition game by integrating the above two adversarial models together can help project two copies of the same vertices with high-dimensional inconsistent features into the same low-dimensional embedding space, and thus guarantee the distribution consistency between two networks in terms of both distance and shape.

In this paper, we propose a novel Heterogeneous Gaussian Mechanism (HGM) to preserve differential privacy in deep neural networks, with provable robustness against adversarial examples. We first relax the constraint of the privacy budget in the traditional Gaussian Mechanism from (0, 1] to (0, infty), with a new bound of the noise scale to preserve differential privacy. The noise in our mechanism can be arbitrarily redistributed, offering a distinctive ability to address the trade-off between model utility and privacy loss. To derive provable robustness, our HGM is applied to inject Gaussian noise into the first hidden layer. Then, a tighter robustness bound is proposed. Theoretical analysis and thorough evaluations show that our mechanism notably improves the robustness of differentially private deep neural networks, compared with baseline approaches, under a variety of model attacks.

Relation Extraction (RE) is one of the fundamental tasks in Information Extraction and Natural Language Processing. Dependency trees have been shown to be a very useful source of information for this task. The current deep learning models for relation extraction has mainly exploited this dependency information by guiding their computation along the structures of the dependency trees. One potential problem with this approach is it might prevent the models from capturing important context information beyond syntactic structures and cause the poor cross-domain generalization. This paper introduces a novel method to use dependency trees in RE for deep learning models that jointly predicts dependency and semantics relations. We also propose a new mechanism to control the information flow in the model based on the input entity mentions. Our extensive experiments on benchmark datasets show that the proposed model outperforms the existing methods for RE significantly.

Network embedding aims to learn latent low dimensional representation of vertices in graphs while preserving the intrinsic characteristics of graph data. In this paper, we propose a density-aware local autoencoder embedding architecture, DAL-SAE, with three features. First, we develop a flexible density-aware local deep autoencoder embedding method to perform local embedding on each of K clustering-based subgraphs with the optimization at both vertex and subgraph levels, in response to imbalanced density-based local characteristics of vertices and subgraphs. We design K local autoencoder embedding models, each with individual parameters and structure, to jointly train K subgraphs and optimize the loss functions within and across clusters. Second, we design an autoencoder graph clustering method to optimize local embedding and graph clustering simultaneously and capture local, clustering, and global network structure in the learning process. Third but last, a density-aware local Siamese autoencoder embedding approach can be utilized to train multiple clustering-based subgraphs with similar local characteristics on the common Siamese networks, to save the memory consumption of multiple local embedding models as well as maintain the similar embedding features.

Evaluating on adversarial examples has become a standard procedure to measure robustness of deep learning models. Due to the difficulty of creating white-box adversarial examples for discrete text input, most analyses of the robustness of NLP models have been done through black-box adversarial examples. We investigate adversarial examples for character-level neural machine translation (NMT), and contrast black-box adversaries with a novel white-box adversary, which employs differentiable string-edit operations to rank adversarial changes. We propose two novel types of attacks which aim to remove or change a word in a translation, rather than simply break the NMT. We demonstrate that white-box adversarial examples are significantly stronger than their black-box counterparts in different attack scenarios, which show more serious vulnerabilities than previously known. In addition, after performing adversarial training, which takes only 3 times longer than regular training, we can improve the model’s robustness significantly.

Traditional network representation learning techniques aim to learn latent low-dimensional representation of vertices in graphs. This paper presents a novel edge representation learning framework, GANDLERL, that combines generative adversarial network based multi-label classification with density-adaptive local edge representation learning for producing high-quality low-dimensional edge representations. First, we design a generative adversarial network based multi-label edge classification model to classify rarely labeled edges in graphs with a large amount of noise data into K classes. A four-player zero-sum game model, with the mixed training of true and real-looking fake edges as well as a contrastive loss containing a similar-loss and a dissimilar-loss, is proposed to improve the classification quality of unlabeled edges. Second, a local autoencoder edge representation learning method is developed to design K local representation learning models, each with individual parameters and structure to perform local representation learning on each of K classification-based subgraphs with unique local characteristics and jointly optimize the loss functions within and across classes. Third but last, we propose a density-adaptive edge representation learning method with the optimization at both edge and subgraph levels to address the representation learning of graph data with highly imbalanced vertex degree and edge distribution.

A various number of graph clustering algorithms have been proposed and applied in real-world applications such as network analysis, bio-informatics, social computing, and etc. However, existing algorithms usually focus on optimizing specified quality measures at the global network level, without carefully considering the destruction of local structures which could be informative and significant in practice. In this paper, we propose a novel clustering algorithm for undirected graphs based on a new structure similarity measure which is computed in a recursive procedure. Our method can provide robust and high-quality clustering results, while preserving informative local structures in the original graph. Rigorous experiments conducted on a variety of benchmark and protein datasets show that our algorithm consistently outperforms existing algorithms.

Question retrieval, which aims to find similar versions of a given question, is playing a pivotal role in various question answering (QA) systems. This task is quite challenging, mainly in regard to five aspects: synonymy, polysemy, word order, question length, and data sparsity. In this article, we propose a unified framework to simultaneously handle these five problems. We use the word combined with corresponding concept information to handle the synonymy problem and the polysemous problem. Concept embedding and word embedding are learned at the same time from both the context-dependent and context-independent views. To handle the word-order problem, we propose a high-level feature-embedded convolutional semantic model to learn question embedding by inputting concept embedding and word embedding. Due to the fact that the lengths of some questions are long, we propose a value-based convolutional attentional method to enhance the proposed high-level feature-embedded convolutional semantic model in learning the key parts of the question and the answer. The proposed high-level feature-embedded convolutional semantic model nicely represents the hierarchical structures of word information and concept information in sentences with their layer-by-layer convolution and pooling. Finally, to resolve data sparsity, we propose using the multi-view learning method to train the attention-based convolutional semantic model on question–answer pairs. To the best of our knowledge, we are the first to propose simultaneously handling the above five problems in question retrieval using one framework. Experiments on three real question-answering datasets show that the proposed framework significantly outperforms the state-of-the-art solutions.

Rumor stance classification is the task of determining the stance towards a rumor in text. This is the first step in effective rumor tracking on social media which is an increasingly important task. In this work, we analyze Twitter users' stance toward a rumorous tweet, in which users could support, deny, query, or comment upon the rumor. We propose a deep attentional CNN-LSTM approach, which takes the sequence of tweets in a thread of conversation as the input. We use neighboring tweets in the timeline as context vectors to capture the temporal dynamism in users' stance evolution. In addition, we use extra features such as friendship, to leverage useful relational features that are readily available in social media. Our model achieves the state-of-the-art results on rumor stance classification on a recent SemEval dataset, improving accuracy and F1 score by 3.6% and 4.2% respectively.

In this paper, we focus on developing a novel mechanism to preserve differential privacy in deep neural networks, such that: (1) The privacy budget consumption is totally independent of the number of training steps; (2) It has the ability to adaptively inject noise into features based on the contribution of each to the output; and (3) It could be applied in a variety of different deep neural networks. To achieve this, we figure out a way to perturb affine transformations of neurons, and loss functions used in deep neural networks. In addition, our mechanism intentionally adds "more noise" into features which are "less relevant" to the model output, and vice-versa. Our theoretical analysis further derives the sensitivities and error bounds of our mechanism. Rigorous experiments conducted on MNIST and CIFAR-10 datasets show that our mechanism is highly effective and outperforms existing solutions.

Human behavior prediction is critical in understanding and addressing large scale health and social issues in online communities. Specifically, predicting when in the future a user will engage in a behavior as opposed to whether a user will behave at a particular time is a less studied subproblem of behavior prediction. Further lacking is exploration of how social context affects personal behavior and the exploitation of network structure information in behavior and time prediction. To address these problems we propose a novel semi-supervised deep learning model for prediction of return time to personal behavior. A carefully designed objective function ensures the model learns good social context embeddings and historical behavior embeddings in order to capture the effects of social influence on personal behavior. Our model is validated on a unique health social network dataset by predicting when users will engage in physical activities. We show our model outperforms relevant time prediction baselines.

In recent years, deep learning has spread beyond both academia and industry with many exciting real-world applications. The development of deep learning has presented obvious privacy issues. However, there has been lack of scientific study about privacy preservation in deep learning. In this paper, we concentrate on the auto-encoder, a fundamental component in deep learning, and propose the deep private auto-encoder (dPA). Our main idea is to enforce ε-differential privacy by perturbing the objective functions of the traditional deep auto-encoder, rather than its results. We apply the dPA to human behavior prediction in a health social network. Theoretical analysis and thorough experimental evaluations show that the dPA is highly effective and efficient, and it significantly outperforms existing solutions.

Chronic obstructive pulmonary disease (COPD) and lung cancer (LC) are two serious diseases that present a major health problem worldwide. However, genetic contribution to both diseases remains unclear, including various regulation mechanisms at genetic level resulting in the progression from COPD to LC. In this paper, we describe our comprehensive methodologies, which seamlessly integrate both biological (conducted in “wet labs”) and computational (based on domain ontologies and semantic technologies) approaches, to investigate the important role of microRNA::mRNA regulations performed in COPD and LC. We discovered two genes, RGS6 and PARK2, that are strongly associated with the risk of developing either COPD or LC or both; additionally, we also identified two sets of microRNAs that are computationally predicted to regulate RGS6 and PARK2, respectively. These microRNAs can be further biologically verified in the future and serve as novel biomarkers in COPD and/or LC.

Distributed word representations are able to capture syntactic and semantic regularities in text. In this paper, we present a word representation scheme that incorporates authorship information. While maintaining similarity among related words in the induced distributed space, our word vectors can be effectively used for some text classification tasks too. We build on a log-bilinear document model (lbDm), which extracts document features, and word vectors based on word co-occurrence counts. First, we propose a log-bilinear author model (lbAm), which contains an additional author matrix. We show that by directly learning author feature vectors, as opposed to document vectors, we can learn better word representations for the authorship attribution task. Furthermore, authorship information has been found to be useful for sentiment classification. We enrich the author model with a sentiment tensor, and demonstrate the effectiveness of this hybrid model (lbHm) through our experiments on a movie review-classification dataset.

Classifying the stance expressed in online microblogging social media is an emerging problem in opinion mining. We propose a probabilistic approach to stance classification in tweets, which models stance, target of stance, and sentiment of tweet, jointly. Instead of simply conjoining the sentiment or target variables as extra variables to the feature space, we use a novel formulation to incorporate three-way interactions among sentiment-stance-input variables and three-way interactions among target-stance-input variables. The proposed specification intuitively aims to discriminate sentiment features from target features for stance classification. In addition, regularizing a single stance classifier, which handles all targets, acts as a soft weight-sharing among them. We demonstrate that discriminative training of this model achieves the state-of-the-art results in supervised stance classification, and its generative training obtains competitive results in the weakly supervised setting.

Supervised stance classification, in such domains as Congressional debates and online forums, has been a topic of interest in the past decade.Approaches have evolved from text classification to structured output prediction, including collective classification and sequence labeling.In this work, we investigate collective classification of stances on Twitter, using hinge-loss Markov random fields (HL-MRFs).Given the graph of all posts, users, and their relationships, we constrain the predicted post labels and latent user labels to correspond with the network structure.We focus on a weakly supervised setting, in which only a small set of hashtags or phrases is labeled.Using our relational approach, we are able to go beyond the stance-indicative patterns and harvest more stance-indicative tweets, which can also be used to train any linear text classifier when the network structure is not available or is costly.

Human behavior prediction is critical to studying how healthy behavior can spread through a social network. In this work we present a novel user representation based human behavior prediction model, the User Representation-based Socialized Gaussian Process model (UrSGP). First, we present the Deep Interaction Representation Learning (Deep Interaction) model for learning latent representations of interaction social networks in which each user is characterized by a set of attributes. In particular, we consider social interaction factors and user attribute factors to build a bimodal, fixed representation of each user in the network. Our model aims to capture the evolution of social interactions and user attributes and learn the hidden correlations between them. We then use our latent features for human behavior prediction via the UrSGP model. An empirical experiment conducted on a real health social network demonstrates that our model outperforms baseline approaches for human behavior prediction.

Modeling physical activity propagation, such as activity level and intensity, is a key to preventing obesity from cascading through communities, and to helping spread wellness and healthy behavior in a social network. However, there have not been enough scientific and quantitative studies to elucidate how social communication may deliver physical activity interventions. In this work, we introduce a novel model named T opic-aware C ommunity-level P hysical Activity Propagation with T emporal Dynamics (TCPT) to analyze physical activity propagation and social influence at different granularities (i.e., individual level and community level). Given a social network, the TCPT model first integrates the correlations between the content of social communication, social influences, and temporal dynamics. Then, a hierarchical approach is utilized to detect a set of communities and their reciprocal influence strength of physical activities. The experimental evaluation shows not only the effectiveness of our approach but also the correlation of the detected communities with various health outcome measures. Our promising results pave a way for knowledge discovery in health social networks.

Modeling and predicting human behaviors, such as the activity level and intensity, is the key to prevent the cascades of obesity, and help spread wellness and healthy behavior in a social network. The user diversity, dynamic behaviors, and hidden social influences make the problem more challenging. In this work, we propose a deep learning model named Social Restricted Boltzmann Machine (SRBM) for human behavior modeling and prediction in health social networks. In the proposed SRBM model, we naturally incorporate self-motivation, implicit and explicit social influences, and environmental events together into three layers which are historical, visible, and hidden layers. The interactions among these behavior determinants are naturally simulated through parameters connecting these layers together. The contrastive divergence and back-propagation algorithms are employed for training the model. A comprehensive experiment on real and synthetic data has shown the great effectiveness of our deep learning model compared with conventional methods.

Identification of non-coding RNAs (ncRNAs) has been significantly enhanced due to the rapid advancement in sequencing technologies. On the other hand, semantic annotation of ncRNA data lag behind their identification, and there is a great need to effectively integrate discovery from relevant communities. To this end, the Non-Coding RNA Ontology (NCRO) is being developed to provide a precisely defined ncRNA controlled vocabulary, which can fill a specific and highly needed niche in unification of ncRNA biology.

Research has indicated that microRNAs (miRNAs), a special class of non-coding RNAs (ncRNAs), can perform important roles in different biological and pathological processes. miRNAs' functions are realized by regulating their respective target genes (targets). It is thus critical to identify and analyze miRNA-target interactions for a better understanding and delineation of miRNAs' functions. However, conventional knowledge discovery and acquisition methods have many limitations. Fortunately, semantic technologies that are based on domain ontologies can render great assistance in this regard. In our previous investigations, we developed a miRNA domain-specific application ontology, Ontology for MIcroRNA Target (OMIT), to provide the community with common data elements and data exchange standards in the miRNA research. This paper describes (1) our continuing efforts in the OMIT ontology development and (2) the application of the OMIT to enable a semantic approach for knowledge capture of miRNA-target interactions.

We present a novel approach for relation classification, using a recursive neural network (RNN), based on the shortest path between two entities in a dependency graph.Previous works on RNN are based on constituencybased parsing because phrasal nodes in a parse tree can capture compositionality in a sentence.Compared with constituency-based parse trees, dependency graphs can represent relations more compactly.This is particularly important in sentences with distant entities, where the parse tree spans words that are not relevant to the relation.In such cases RNN cannot be trained effectively in a timely manner.However, due to the lack of phrasal nodes in dependency graphs, application of RNN is not straightforward.In order to tackle this problem, we utilize dependency constituent units called chains.Our experiments on two relation classification datasets show that Chain based RNN provides a shallower network, which performs considerably faster and achieves better classification results.

Modeling physical activity propagation, such as the activity level and intensity, is the key to prevent the cascades of obesity, and help spread wellness and healthy behavior in a social network. However, there has been lacking of scientific and quantitative study to elucidate how social communication may deliver physical activity interventions. In this work we introduce a Community-level Physical Activity Propagation (CPP) model to analyze physical activity propagation and social influence at different granularities (i.e., individual level and community level). CPP is a novel model which is inspired by the well-known Independent Cascade and Community-level Social Influence models. Given a social network, we utilize a hierarchical approach to detect a set of communities and their reciprocal influence strength of physical activities. CPP provides a powerful tool to discover, summarize, and investigate influence patterns of physical activities in a health social network. The detail experimental evaluation shows not only the effectiveness of our approach but also the correlation of the detected communities with various health outcome measures (i.e., both existing ones and our novel measure, named Wellness score, which is a combination of lifestyle parameters, biometrics, and biomarkers). Our promising results potentially pave a way for knowledge discovery in health social networks.

Although automatic text grading systems have reached an accuracy level comparable to human grading, with successful commercial and research implementations (e.g., Latent Semantic Analysis), these systems can provide limited feedback about which statements of the text are incorrect and why they are incorrect. In the present work, we propose the use of a hybrid Ontology-based Information Extraction (OBIE) system to identify both correct and incorrect statements by combining extraction rules and machine learning based information extractors. Experiments show that given 77 student answers to a Cell Biology final exam question, our hybrid system can identify both correct and incorrect statements with high precision and recall measures.

As researchers analyze huge amounts of data that are annotated by large biomedical ontologies, one of the major challenges for data mining and machine learning is to leverage both ontologies and data together in a systematic and scalable way. In this paper, we address two interesting and related problems for mining biomedical ontologies and data: i) how to discover semantic associations with the help of formal ontologies, ii) how to identify potential errors in the ontologies with the help of data. By representing both ontologies and data using RDF hyper graphs, and subsequently transforming the hyper graphs to corresponding bipartite forms, we provide a generalized data mining method that scales beyond what existing ontology-based approaches can provide. We show the proposed method is indeed capable of capturing semantic associations while seamlessly incorporate domain knowledge in ontologies by performing evaluations on real-world electronic health dataset and NCBO ontologies. We also show that our data mining methods can discover and suggest corrections for misinformation in biomedical ontologies.

Automatic grading systems for summaries and essays have been studied for years. Most commercial and research implementations are based in statistical methods, such as Latent Semantic Analysis (LSA), which can provide high accuracy on similarity between the essay and the graded or standard essays, but they can offer very limited feedback. In the present work, we propose a novel method to provide both grades and meaningful feedback for student summaries by Ontology-based Information Extraction (OBIE). We use ontological concepts and relationships to create extraction rules to identify correct statements. Based on ontology constraints (e.g., disjointness between concepts), we define patterns that are logically inconsistent with the ontology to create rules to extract incorrect statements. Experiments show that the grades given to 18 student summaries on Ecosystems by OBIE are correlated to human gradings. OBIE also provide meaningful feedback on the errors those students made in their summaries.

A number of text mining and information extraction projects such as Text Runner and NELL seek to automatically build knowledge bases from the rapidly growing amount of information on the web. In order to scale to the size of the web, these projects often employ ad hoc heuristics to reason about uncertain and contradictory information rather than reasoning jointly about all candidate facts. In this paper, we present a Markov logic-based system for cleaning an extracted knowledge base. This allows a scalable system such as NELL to take advantage of joint probabilistic inference, or, conversely, allows Markov logic to be applied to a web scale problem. Our system uses only the ontological constraints and confidence values of the original system, along with human-labeled data if available. The labeled data can be used to calibrate the confidence scores from the original system or learn the effectiveness of individual extraction patterns. To achieve scalability, we introduce a neighborhood grounding method that only instantiates the part of the network most relevant to the given query. This allows us to partition the knowledge cleaning task into tractable pieces that can be solved individually. In experiments on NELL's knowledge base, we evaluate several variants of our approach and find that they improve both F1 and area under the precision-recall curve.

Modeling and predicting human behaviors, such as the activity level and intensity, is the key to prevent the cascades of obesity, and help spread wellness and healthy behavior in a social network. In this work, we propose a Socialized Gaussian Process (SGP) for socialized human behavior modeling. In the proposed SGP model, we naturally incorporates human's personal behavior factor and social correlation factor into a unified model, where basic Gaussian Process model is leveraged to capture individual's personal behavior pattern. Furthermore, we extend the Gaussian Process Model to socialized Gaussian Process (SGP) which aims to capture social correlation phenomena in the social network. The detailed experimental evaluation has shown the SGP model achieves the best prediction accuracy compared with other baseline methods.

Naive Bayesian learning has been popular in data mining applications. However, the performance of naive Bayesian learning is sometimes poor due to the unrealistic assumption that all features are equally important and independent given the class value. Therefore, it is widely known that the performance of naive Bayesian learning can be improved by mitigating this assumption, and many enhancements to the basic naive Bayesian learning have been proposed to resolve this problem including feature selection and feature weighting. In this paper, we propose a new method for calculating the weights of features in naive Bayesian learning using Kullback-Leibler measure. Empirical results are presented comparing this new feature weighting method with some other methods for a number of datasets.

In this paper, we address an interesting data mining problem of finding semantically associated itemsets, i.e., items connected via indirect links. We propose a novel method for discovering semantically associated itemsets based on a hypergraph representation of the database. We describe two similarity measures to compute the strength of associations between items. Specifically, we introduce the average commute time similarity, sCT, based on the random walk model on hypergraph, and the inner-product similarity, sL+, based on the Moore-Penrose pseudoinverse of the hypergraph Laplacian matrix. Given semantically associated 2-itemsets generated by these measures, we design a hypergraph expansion method with two search strategies, namely, the clique and connected component search, to generate k-itemsets (k >; 2). We show the proposed method is indeed capable of capturing semantically associated itemsets through experiments performed on three datasets ranging from low to high dimensionality. The semantically associated itemsets discovered in our experiment is promising to provide valuable insights on interrelationship between medical concepts and other domain specific concepts.

Information Extraction (IE) has existed as a field for several decades and has produced some impressive systems in the recent past. Despite its success, widespread usage and commercialization remain elusive goals for this field. We identify the lack of effective mechanisms for reuse as one major reason behind this situation. Here, we mean not only the reuse of the same IE technique in different situations but also the reuse of information related to the application of IE techniques (e.g., features used for classification).

Financial forecasting is the basis for budgeting activities and estimating future financing needs. Applying machine learning and data mining models to financial forecasting is both effective and efficient. Among different kinds of machine learning models, kernel methods are well accepted since they are more robust and accurate than traditional models, such as neural networks. However, learning from multiple data sources is still one of the main challenges in the financial forecasting area. In this paper, we focus on applying the multiple kernel learning models to the multiple major international stock indexes. Our experiment results indicate that applying multiple kernel learning to the financial forecasting problem suffers from both the short training period problem and non-stationary problem. Therefore we propose a novel multiple kernel learning model to address the challenge by introducing the Gompertz model and considering a non-linear combination of different kernel matrices. The experiment results show that our Gompertz multiple kernel learning model addresses the challenges and achieves better performance than the original multiple kernel learning model and single SVM models.

Ontology-Based Information Extraction (OBIE) has recently emerged as a subfield of Information Extraction (IE). Here, ontologies- which provide formal and explicit specifications of conceptualizations- play a crucial role in the information extraction process. Several OBIE systems have been implemented previously but all of them use a single ontology although multiple ontologies have been designed for many domains. We have studied the theoretical basis for using multiple ontologies in information extraction and have developed information extraction systems that use them. These systems investigate the two major scenarios for having multiple ontologies for the same domain: specializing in subdomains and providing different perspectives. The domain of universities has been used for the former scenario through a corpus collected from university websites. For the latter, the domain of terrorist attacks and a corpus used by a previous Message Understanding Conference (MUC) have been used. The results from these two case studies indicate that using multiple ontologies in information extraction has led to a clear improvement in performance measures.

Event-related potentials (ERP) are brain electrophysiological patterns created by averaging electroencephalographic (EEG) data, time-locking to events of interest (e.g., stimulus or response onset). In this paper, we propose a generic framework for mining anddeveloping domain ontologies and apply it to mine brainwave (ERP) ontologies. The concepts and relationships in ERP ontologies can be mined according to the following steps: pattern decomposition, extraction of summary metrics for concept candidates, hierarchical clustering of patterns for classes and class taxonomies, and clustering-based classification and association rules mining for relationships (axioms) of concepts. We have applied this process to several dense-array (128-channel) ERP datasets. Results suggest good correspondence between mined concepts and rules, on the one hand, and patterns and rules that were independently formulated by domain experts, on the other. Data mining results also suggest ways in which expert-defined rules might be refined to improve ontologyrepresentation and classification results. The next goal of our ERP ontology mining framework is to address some long-standing challenges in conducting large-scale comparison and integration of results across ERP paradigms and laboratories. In a more general context, this work illustrates the promise of an interdisciplinary research program, which combines data mining, neuroinformatics andontology engineering to address real-world problems.

Abnormal events, such as security attacks, misconfigurations, or electricity failures, could have severe consequences toward the normal operation of the Border Gateway Protocol (BGP) that is in charge of the delivery of packets between different autonomous domains, a key operation for the Internet to function. Unfortunately, it has been a difficult task for network security researchers and engineers to classify and detect these events. In our previous work, we have shown that with classification (which relies on the labeling with domain knowledge from BGP experts), it is feasible to effectively detect and distinguish some worms and blackouts from normal BGP behaviors. In this paper, we move one important step forward—we show that we can automatically detect and classify between different abnormal BGP events based on a hierarchy discovered by clustering. As a systematic application of data mining, we devise a clustering method based on normalized BGP data that forms a tree-like hierarchy of abnormal BGP event classes. We then obtain a set of classification rules for each class (node) in the hierarchy, thus able to label unknown BGP data to a closest class. Our method works even as the BGP dynamics evolve over time, as shown in our experiments with seven different abnormal events during a four-year period. Our work, in a more general context, shows it is promising to conduct an interdisciplinary research between network security and data mining in solving real-world problems.