Medical Large Vision-Language Models (Med-LVLMs) have shown promising results in clinical applications, but often suffer from hallucinated outputs due to misaligned visual understanding. In this work, we identify two fundamental limitations contributing to this issue: insufficient visual representation learning and poor visual attention alignment. To address these problems, we propose MedAlign, a simple, lightweight alignment distillation framework that transfers visual alignment knowledge from a domain-specific Contrastive Language-Image Pre-training (CLIP) model to Med-LVLMs. MedAlign introduces two distillation losses: a spatial-aware visual alignment loss based on visual token-level similarity structures, and an attention-aware distillation loss that guides attention toward diagnostically relevant regions. Extensive experiments on medical report generation and medical visual question answering (VQA) benchmarks show that MedAlign consistently improves both performance and interpretability, yielding more visually grounded outputs.

Knowledge Graph (KG)-supported Graph Neural Network models are becoming crucial in recommendation systems due to their ability to mitigate the data sparsity challenge. However, these models remain suboptimal because they overlook the representation differences between the inherent user-item Bipartite Graph (BG) and the external head-relation-tail KG, leading to semantic misalignment. Moreover, they indiscriminately incorporate various types of relations from the KG, which may introduce noise information into the model, ultimately degrading recommendation performance. To address these challenges, we propose an end-to-end model named Multi-graph Fusion Cross-model Contrastive Learning (MFCCL). To uncover users' interest in items and explore the associations between items, we first construct a user-interest graph by integrating information from both the BG and KG, and an item-association graph derived from the BG. We devise a multi-graph representation learning module that incorporates rich semantics into user and item representations in parallel. Simultaneously, a classical collaborative filtering module is introduced to fully leverage user-item collaborative signals. Additionally, we design a novel free data-augmentation cross-model contrastive learning to facilitate the exchange of complementary information between different models. Empirical evaluations on three widely used benchmarks demonstrate that our MFCCL method achieved significant improvements over the baselines.

Multi-graph multi-label learning (MGML) represents each object as a bag-of-graphs with multiple labels, but demands large-scale labeled data whose acquisition is often difficult and costly. Self-supervised contrastive learning (SCL) mitigates label dependence by leveraging data augmentation to construct discriminative pretext tasks, proving effective for multi-instance learning. However, when applied to MGML, SCL faces two key challenges: (1) it distinguishes individual instances by their differences, whereas MGML requires modeling label correlations; (2) it assumes semantic invariance under augmentation, but structural perturbations in MGML alter label semantics. To tackle these challenges, we propose a self-suPervised contrastive rE-learning framework for mulTi-grAph multi-labeL classification (PETAL). Specifically, to model label correlations, we first define a unified label space to learn label prototypes and align features with them, yielding prototype-aligned representations. We then design a multi-granularity contrastive loss over these representations, which captures label dependencies by contrasting at the bag level, graph level, and bag-graph level. Moreover, to ensure semantic invariance, we develop a contrastive re-learning strategy based on prototype-aligned representations to generate augmentation-free positive samples. This guarantees consistent multi-label distributions without structural perturbations. Experiments on six datasets demonstrate that PETAL achieves an average improvement of 4.12% over state-of-the-art self-supervised and supervised baselines.

Large Language Models (LLMs) are increasingly integrated into various domains, making knowledge editing techniques crucial yet potentially hazardous.Current editing methods primarily target atomic facts, overlooking the significant risks associated with manipulating "factual opinions", e.g., documented stances of public figures on societal issues.Such manipulation could reshape public images, influence elections, and alter societal views.To systematically assess this threat, we introduce the Factual Opinion Editing with Evidence (FOE) benchmark, which encompasses 261 public figures, 19 issue categories, and 2,178 complete opinion records.Our evaluations demonstrate that current editing techniques struggle significantly with factual opinions, often achieving only superficial changes while failing to preserve consistency between the edited opinion and the supporting evidence generated by the model.To address this limitation, we further propose a simple yet effective Self-Generated Evidence-Aligned method that achieves opinion-evidence alignment without relying on explicit instructions.Together, our benchmark and method provide a foundation for understanding the emerging security implications of factual opinion editing in LLMs.

This paper presents AutoRAN 1 , the first framework to automate the hijacking of internal safety reasoning in large reasoning models (LRMs).At its core, AutoRAN pioneers an execution simulation paradigm that leverages a weaker but less-aligned model to simulate execution reasoning for initial hijacking attempts and iteratively refine attacks by exploiting reasoning patterns leaked through the target LRM's refusals.This approach steers the target model to bypass its own safety guardrails and elaborate on harmful instructions.We evaluate AutoRAN against state-of-the-art LRMs, including gpt-o3/o4-mini and Gemini-2.5-Flash,across multiple benchmarks (AdvBench, Harm-Bench, and StrongReject).Results show that AutoRAN achieves approaching 100% success rate within one or a few turns, effectively neutralizing reasoning-based defenses even when evaluated by robustly aligned external models.This work reveals that the transparency of the reasoning process itself creates a critical and exploitable attack surface, highlighting the urgent need for new defenses that protect models' reasoning traces rather than merely their final outputs.The code for replicating Au-toRAN is available at: https://github.com/ JACKPURCELL/AutoRAN-public.

The explainable medical coding task aims to automatically assign International Classification of Diseases (ICD) codes to clinical notes while providing explicit justifications for each assignment.Recent approaches employ large language models (LLMs) to generate such explanations.However, their performance remains limited due to a lack of understanding of the clinical meanings of ICD codes.Additionally, the vast ICD code space further complicates the task of accurate prediction.To address these challenges, we propose the ICDAGENT framework, which consists of two collaborative LLM agents: a coding agent and a critical agent.The coding agent extracts ICD codes and generates preliminary rationales, while the critical agent performs fine-grained chain-of-thought reasoning to verify and refine them.Furthermore, the critical agent is trained with a rationaleaware reward, combined with reinforcement learning, enabling it to distinguish between correct and incorrect reasoning and ensure explanation accuracy.Experiments across multiple ICD coding standards and datasets demonstrate that ICDAGENT achieves effective ICD coding with accurate and trustworthy explanations.1

Multi-Graph Learning (MGL) is a fundamental machine learning paradigm that represents objects as bags-of-graphs, each encoding a distinct structural property, and has broad applications in bioinformatics, chemistry, computing power networks, and software defect detection. However, the inherent scarcity of labeled data poses a significant bottleneck for supervised MGL approaches. While self-supervised contrastive learning offers a compelling solution, its direct application to MGL faces three key challenges: (1) existing graph neural networks, primarily for single-graph modeling, struggle to yield discriminative bag-level representations from bags-of-graphs; (2) conventional contrastive objectives are limited to single-level settings, failing to capture cross-hierarchical dependencies; and (3) standard data augmentation often disrupts intrinsic graph and bag structures, undermining semantic consistency. To address these issues, we propose the Hierarchical Graph-Bag-Network (HGBN), a self-supervised MGL framework that constructs hierarchical representations in the form of a graph-bag-network. HGBN employs an asymmetric hierarchical graph neural network to learn discriminative graph-level and bag-level representations, introduces cross-hierarchical contrastive objectives to align graph-level and bag-level semantics, and leverages the asymmetric network outputs to form positive and negative pairs, preserving intrinsic structural and semantic consistency. Experiments on eight benchmark multi-graph datasets demonstrate that HGBN consistently outperforms both supervised and self-supervised state-of-the-art baselines, achieving average improvements of 4.82% in accuracy and F1 score.

Multi-label few-shot image classification is a crucial and challenging task due to limited annotated data and elusive category specificity. However, research on this topic is still in the rudimentary stage and few methods are available. Existing methods either leverage data augmentation to alleviate data scarcity or utilize label features as auxiliary knowledge to eliminate the negative effect caused by irrelevant categories, but they ignore the utilization of image region features for data augmentation, and overlook to learn appropriate text feature to better match the image features of specific categories. Moreover, these methods only focus on one side and do not effectively tackle the above two issues simultaneously. In this paper, we introduce a novel prototype-based multi-label few-shot learning framework that seamlessly integrates pairwise feature augmentation and flexible prompt learning. Specifically, by pairwise feature augmentation, we leverage the region features of images in the support set to generate more image features and construct image prototypes, thus alleviating the issue of data scarcity. By flexible prompt learning, we adaptively acquire class-specific prompts to build text prototypes that highly match the image features of specific classes, thereby mitigating the impact of irrelevant classes. Finally, with adaptive learnable parameters, we merge image and text prototypes to obtain the final prototypes, achieving a more powerful classifier for multi-label few-shot image classification. Extensive experimental results demonstrate that our proposed method can push the performance to a higher level.

With the continuously expanding parameters, efficiently adapting large language models to downstream tasks is crucial in resource-limited conditions.Many parameter-efficient finetuning methods have emerged to address this challenge.However, they lack flexibility, like LoRA requires manually selecting trainable parameters and rank size, (IA) 3 can only scale the activations along columns, yielding inferior results due to less precise fine-tuning.To address these issues, we propose a novel method named AdaDHP with fewer parameters and finer granularity, which can adaptively select important parameters for each task.Specifically, we introduce two trainable vectors for each parameter and fine-tune the parameters through Hadamard product along both rows and columns.This significantly reduces the number of trainable parameters, with our parameter count capped at the lower limit of LoRA.Moreover, we design an adaptive parameter selection strategy to select important parameters for downstream tasks dynamically.This allows our method to flexibly remove unimportant parameters for downstream tasks.Finally, we demonstrate the superiority of our method on the T5-base model across 17 NLU tasks and on complex mathematical tasks with the Llama series models.

Aofei Chang, Le Huang, Alex James Boyd, Parminder Bhatia, Taha Kass-Hout, Cao Xiao, Fenglong Ma. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Differentially private diffusion models (DPDMs) harness the remarkable generative capabilities of diffusion models while enforcing differential privacy (DP) for sensitive data. However, existing DPDM training approaches often suffer from significant utility loss, large memory footprint, and expensive inference cost, impeding their practical uses. To overcome such limitations, we present RAPID: Retrieval Augmented PrIvate Diffusion model, a novel approach that integrates retrieval augmented generation (RAG) into DPDM training. Specifically, RAPID leverages available public data to build a knowledge base of sample trajectories; when training the diffusion model on private data, RAPID computes the early sampling steps as queries, retrieves similar trajectories from the knowledge base as surrogates, and focuses on training the later sampling steps in a differentially private manner. Extensive evaluation using benchmark datasets and models demonstrates that, with the same privacy guarantee, RAPID significantly outperforms state-of-the-art approaches by large margins in generative quality, memory footprint, and inference cost, suggesting that retrieval-augmented DP training represents a promising direction for developing future privacy-preserving generative models. The code is available at: https://github.com/TanqiuJiang/RAPID

Vision foundation models (FMs) like CLIP have exhibited exceptional capabilities in visual and linguistic understanding, particularly in zero-shot inference tasks. However, these models struggle with data that significantly deviates from their training samples, necessitating fine-tuning, which is often infeasible in centralized settings due to data privacy concerns. Federated learning (FL) combined with parameter-efficient fine-tuning (PEFT) offers a potential solution, yet existing methods face issues with domain-specific characteristics and out-of-domain generalization. We propose a cross-silo Federated Adapter Generalization (FedAG), a novel federated fine-tuning approach that leverages multiple fine-grained adapters to capture domain-specific knowledge while enhancing out-of-domain generalization. Our method uses quality-aware in-domain mutual learning and attention-regularized cross-domain learning to integrate domain-specific insights effectively. Experiments of the CLIP model on three domain-shifting datasets, ImageCLEF-DA, Office-Home, and DomainNet, demonstrate the superior performance of FedAG in both in-domain and out-of-domain scenarios. We envision this work as a milestone for generalizing CLIP to handle the challenge of out-of-domain knowledge under federated learning setting.

Diffusion models have recently emerged as a state-of-the-art approach for synthetic Electronic Health Record (EHR) generation, offering superior fidelity and diversity over traditional generative models. However, existing diffusion-based methods struggle with unique challenges: limited representation learning and modality utilization, where they fail to explicitly capture inter-modality dependencies and fine-grained code-level interactions, and constrained adaptability due to reliance on U-Net-based architectures, which are not well-suited for handling the heterogeneous and evolving nature of EHR data. Furthermore, current evaluation paradigms rely on either perplexity-based sequence modeling or global distributional measures, lacking robustness in assessing both intra-visit code relationships and inter-visit temporal patterns. To address these limitations, we propose MedDiTPro, a diffusion transformer-based framework that enhances multimodal EHR generation by integrating structured modality-aware guidance. Through a unified transformer for intra-visit representation learning, a modality-specific and datawise prompt learner, and a diffusion transformer with structured guidance, MedDiTPro achieves state-of-the-art performance in generating diverse and clinically meaningful synthetic records. Extensive experiments on publicly available datasets demonstrate that MedDiTPro achieves state-of-the-art fidelity, privacy preservation, and utility.

Geographic health disparities pose a pressing global challenge, particularly in underserved regions of low- and middle-income nations. Addressing this issue requires a collaborative approach to enhance healthcare quality, leveraging support from medically more developed areas. Federated learning emerges as a promising tool for this purpose. However, the scarcity of medical data and limited computation resources in underserved regions make collaborative training of powerful machine learning models challenging. Furthermore, there exists an asymmetrical reciprocity between underserved and developed regions. To overcome these challenges, we propose a novel cross-silo federated learning framework, named FedHelp, aimed at alleviating geographic health disparities and fortifying the diagnostic capabilities of underserved regions. Specifically, FedHelp leverages foundational model knowledge via one-time API access to guide the learning process of underserved small clients, addressing the challenge of insufficient data. Additionally, we introduce a novel asymmetric dual knowledge distillation module to manage the issue of asymmetric reciprocity, facilitating the exchange of necessary knowledge between developed large clients and underserved small clients. We validate the effectiveness and utility of FedHelp through extensive experiments on both medical image classification and segmentation tasks. The experimental results demonstrate significant performance improvement compared to state-of-the-art baselines, particularly benefiting clients in underserved regions.

Large language models (LLMs) have shown remarkable performance in various domains, but they are constrained by massive computational and storage costs. Quantization, an effective technique for compressing models to fit resource-limited devices while preserving generative quality, encompasses two primary methods: quantization aware training (QAT) and post-training quantization (PTQ). QAT involves additional retraining or fine-tuning, thus inevitably resulting in high training cost and making it unsuitable for LLMs. Consequently, PTQ has become the research hotspot in recent quantization methods. However, existing PTQ methods usually rely on various complex computation procedures and suffer from considerable performance degradation under low-bit quantization settings. To alleviate the above issues, we propose a simple and effective post-training quantization paradigm for LLMs, named SEPTQ. Specifically, SEPTQ first calculates the importance score for each element in the weight matrix and determines the quantization locations in a static global manner. Then it utilizes the mask matrix which represents the important locations to quantize and update the associated weights column-by-column until the appropriate quantized weight matrix is obtained. Compared with previous methods, SEPTQ simplifies the post-training quantization procedure into only two steps, and considers the effectiveness and efficiency simultaneously. Experimental results on various datasets across a suite of models ranging from millions to billions in different quantization bit-levels demonstrate that SEPTQ significantly outperforms other strong baselines, especially in low-bit quantization scenarios.

Estimating brain effective connectivity (EC) from functional magnetic resonance imaging (fMRI) data can aid in comprehending the neural mechanisms underlying human behavior and cognition, providing a foundation for disease diagnosis. However, current spatiotemporal attention modules handle temporal and spatial attention separately, extracting temporal and spatial features either sequentially or in parallel. These approach overlooks the inherent spatiotemporal correlations present in real world fMRI data. Additionally, the presence of noise in fMRI data further limits the performance of existing methods. In this paper, we propose a novel brain effective connectivity estimation method based on Fourier spatiotemporal attention (FSTA-EC), which combines Fourier attention and spatiotemporal attention to simultaneously capture inter-series (spatial) dynamics and intra-series (temporal) dependencies from high-noise fMRI data. Specifically, Fourier attention is designed to convert the high-noise fMRI data to frequency domain, and map the denoised fMRI data back to physical domain, and spatiotemporal attention is crafted to simultaneously learn spatiotemporal dynamics. Furthermore, through a series of proofs, we demonstrate that incorporating learnable filters into fast Fourier transform and inverse fast Fourier transform processes is mathematically equivalent to performing cyclic convolution. The experimental results on simulated and real-resting-state fMRI datasets demonstrate that the proposed method exhibits superior performance when compared to state-of-the-art methods. The code is available at https://github.com/XiongWenXww/FSTA.

Collaborative fairness is a crucial challenge in federated learning. However, existing approaches often overlook a practical yet complex form of heterogeneity: imbalanced covariate shift. We provide a theoretical analysis of this setting, which motivates the design of FedAKD (Federated Asynchronous Knowledge Distillation) - a simple yet effective approach that balances accurate prediction with collaborative fairness. FedAKD consists of client and server updates. In the client update, we introduce a novel asynchronous knowledge distillation strategy based on our preliminary analysis, which reveals that while correctly predicted samples exhibit similar feature distributions across clients, incorrectly predicted samples show significant variability. This suggests that imbalanced covariate shift primarily arises from misclassified samples. Leveraging this insight, our approach first applies traditional knowledge distillation to update client models while keeping the global model fixed. Next, we select the correctly predicted high-confidence samples and update the global model using these samples, while keeping the client models fixed. The server update simply aggregates all client models. We further provide a theoretical proof of FedAKD's convergence. Experimental results on both public datasets (FashionMNIST and CIFAR10) and a real-world Electronic Health Records (EHR) dataset demonstrate that FedAKD significantly improves collaborative fairness, enhances predictive accuracy, and fosters client participation, even under highly heterogeneous data distributions.

Graph self-supervised learning aims to learn the intrinsic graph representations from unlabeled data, with broad applicability in areas such as computing networks. Although graph contrastive learning (GCL) has achieved remarkable progress by generating perturbed views via data augmentation and optimizing sample similarity, it performs poorly in heterophilic graph scenarios (where connected nodes are likely to belong to different classes or exhibit dissimilar features). In heterophilic graphs, existing methods typically rely on random or carefully designed augmentation strategies (e.g., edge dropping) for contrastive views. However, such graph structures exhibit intricate edge relationships, where topological perturbations may completely alter the semantics of neighborhoods. Moreover, most methods focus solely on local contrastive signals while neglecting global structural constraints. To address these limitations, inspired by graph coloring, we propose a novel Coloring learning for heterophilic graph Representation framework, CoRep, which: 1) Pioneers a coloring classifier to generate coloring labels, explicitly minimizing the discrepancy between homophilic nodes while maximizing that of heterophilic nodes. A global positive sample set is constructed using multi-hop same-color nodes to capture global semantic consistency. 2) Introduces a learnable edge evaluator to guide the coloring learning dynamically and utilizes the edges' triplet relations to enhance its robustness. 3) Leverages Gumbel-Softmax to differentially discretize color distributions, suppressing noise via a redundancy constraint and enhancing intra-class compactness. Experimental results on 14 benchmark datasets demonstrate that CoRep significantly outperforms current state-of-the-art methods.

Black-box adversarial attack on vision-language pre-trained models is a practical and challenging task, as text and image perturbations need to be considered simultaneously, and only the predicted results are accessible. Research on this problem is in its infancy, and only a handful of methods are available. Nevertheless, existing methods either rely on a complex iterative cross-search strategy, which inevitably consumes numerous queries, or only consider reducing the similarity of positive image-text pairs but ignore that of negative ones, which will also be implicitly diminished, thus inevitably affecting the attack performance. To alleviate the above issues, we propose a simple yet effective framework to generate high-quality adversarial examples on vision-language pre-trained models, named HQA-VLAttack, which consists of text and image attack stages. For text perturbation generation, it leverages the counter-fitting word vector to generate the substitute word set, thus guaranteeing the semantic consistency between the substitute word and the original word. For image perturbation generation, it first initializes the image adversarial example via the layer-importance guided strategy, and then utilizes contrastive learning to optimize the image adversarial perturbation, which ensures that the similarity of positive image-text pairs is decreased while that of negative image-text pairs is increased. In this way, the optimized adversarial images and texts are more likely to retrieve negative examples, thereby enhancing the attack success rate. Experimental results on three benchmark datasets demonstrate that HQA-VLAttack significantly outperforms strong baselines in terms of attack success rate.

Session-based recommendation is gaining increasing attention due to its practical value in predicting the intents of anonymous users based on limited behaviors. Emerging efforts incorporate various side information to alleviate inherent data scarcity issues in this task, leading to impressive performance improvements. The core of side information-driven session-based recommendation is the discovery and utilization of diverse data. In this survey, we provide a comprehensive review of this task from a data-centric perspective. Specifically, this survey commences with a clear formulation of the task. This is followed by a detailed exploration of various benchmarks rich in side information that are pivotal for advancing research in this field. Afterwards, we delve into how different types of side information enhance the task, underscoring data characteristics and utility. Moreover, we discuss the usage of various side information, including data encoding, data injection, and involved techniques. A systematic review of research progress is then presented, with the taxonomy by the types of side information. Finally, we summarize the current limitations and present the future prospects of this vibrant topic.

The rapid advancement of medical dialog generation (MDG) techniques has enabled medical dialog systems (MDSs) to generate high-quality responses rich in medical expertise by integrating diverse medical information. However, they still encounter several challenges, including generic response generation, lack of semantic precision, and imprecise dialog topic extraction. This study aims to design a novel model to address these challenges simultaneously. Correspondingly, we propose the TRL-HMIE model, which represents transformer reinforcement learning (RL) for heterogeneous medical information extraction. In particular, we incorporate GPT-3 from transformer-related models as the reference language model. Our enhancements focused on three key aspects. First, we developed a conversation-topic classifier to precisely categorize conversation topics, supporting the conversation-topic locator module in generating reliable conversation topics. Second, the model employs a multihead attention mechanism to capture crucial information from the dialog context, facilitating the extraction of key dialog information and enhancing the accuracy of heterogeneous information extraction. Finally, the model integrates RL and a reward fusion mechanism, which, combined with its ability to handle multisource information and long dialog contexts, generates optimized rewards for the TRL-HMIE model, encouraging the production of doctor responses with precise semantics and dialog topics. The experimental results demonstrate that the proposed method achieves a 6.07% improvement over the benchmark model on the MedDG and MedDialog datasets. The experimental results demonstrate that the proposed method achieves a 6.07% improvement over the benchmark model on the MedDG and MedDialog datasets.

Visual Question Answering (VQA) is a fundamental task in computer vision and natural language process fields. Although the “pre-training & finetuning” learning paradigm significantly improves the VQA performance, the adversarial robustness of such a learning paradigm has not been explored. In this paper, we delve into a new problem: using a pre-trained multimodal source model to create adversarial image-text pairs and then transferring them to attack the target VQA models. Correspondingly, we propose a novel VQATTACK model, which can iteratively generate both im- age and text perturbations with the designed modules: the large language model (LLM)-enhanced image attack and the cross-modal joint attack module. At each iteration, the LLM-enhanced image attack module first optimizes the latent representation-based loss to generate feature-level image perturbations. Then it incorporates an LLM to further enhance the image perturbations by optimizing the designed masked answer anti-recovery loss. The cross-modal joint attack module will be triggered at a specific iteration, which updates the image and text perturbations sequentially. Notably, the text perturbation updates are based on both the learned gradients in the word embedding space and word synonym-based substitution. Experimental results on two VQA datasets with five validated models demonstrate the effectiveness of the proposed VQATTACK in the transferable attack setting, compared with state-of-the-art baselines. This work reveals a significant blind spot in the “pre-training & fine-tuning” paradigm on VQA tasks. The source code can be found in the link https://github.com/ericyinyzy/VQAttack.

Depression detection is a challenging and crucial task in psychological illness diagnosis. Utilizing online user posts to predict whether a user suffers from depression seems an effective and promising direction. However, existing methods suffer from either poor interpretability brought by the black-box models or underwhelming performance caused by the completely separate two-stage model structure. To alleviate these limitations, we propose a novel capsule network integrated with contrastive learning for depression detection (DeCapsNet). The highlights of DeCapsNet can be summarized as follows. First, it extracts symptom capsules from user posts by leveraging meticulously designed symptom descriptions, and then distills them into class-indicative depression capsules. The overall workflow is in an explicit hierarchical reasoning manner and can be well interpreted by the Patient Health Questionnaire-9 (PHQ9), which is one of the most widely adopted questionnaires for depression diagnosis. Second, it integrates with contrastive learning, which can facilitate the embeddings from the same class to be pulled closer, while simultaneously pushing the embeddings from different classes apart. In addition, by adopting the end-to-end training strategy, it does not necessitate additional data annotation, and mitigates the potential adverse effects from the upstream task to the downstream task. Extensive experiments on three widely-used datasets show that in both within-dataset and cross-dataset scenarios our proposed method outperforms other strong baselines significantly.

Few-shot and zero-shot text classification aim to recognize samples from novel classes with limited labeled samples or no labeled samples at all. While prevailing methods have shown promising performance via transferring knowledge from seen classes to unseen classes, they are still limited by (1) Inherent dissimilarities among classes make the transformation of features learned from seen classes to unseen classes both difficult and inefficient. (2) Rare labeled novel samples usually cannot provide enough supervision signals to enable the model to adjust from the source distribution to the target distribution, especially for complicated scenarios. To alleviate the above issues, we propose a simple and effective strategy for few-shot and zero-shot text classification. We aim to liberate the model from the confines of seen classes, thereby enabling it to predict unseen categories without the necessity of training on seen classes. Specifically, for mining more related unseen category knowledge, we utilize a large pre-trained language model to generate pseudo novel samples, and select the most representative ones as category anchors. After that, we convert the multi-class classification task into a binary classification task and use the similarities of query-anchor pairs for prediction to fully leverage the limited supervision signals. Extensive experiments on six widely used public datasets show that our proposed method can outperform other strong baselines significantly in few-shot and zero-shot tasks, even without using any seen class samples.

Despite the recent progress in deep neural networks (DNNs), it remains challenging to explain the predictions made by DNNs. Existing explanation methods for DNNs mainly focus on post-hoc explanations where another explanatory model is employed to provide explanations. The fact that post-hoc methods can fail to reveal the actual original reasoning process of DNNs raises the need to build DNNs with built-in interpretability. Motivated by this, many self-explaining neural networks have been proposed to generate not only accurate predictions but also clear and intuitive insights into why a particular decision was made. However, existing self-explaining networks are limited in providing distribution-free uncertainty quantification for the two simultaneously generated prediction outcomes (i.e., a sample's final prediction and its corresponding explanations for interpreting that prediction). Importantly, they also fail to establish a connection between the confidence values assigned to the generated explanations in the interpretation layer and those allocated to the final predictions in the ultimate prediction layer. To tackle the aforementioned challenges, in this paper, we design a novel uncertainty modeling framework for self-explaining networks, which not only demonstrates strong distribution-free uncertainty modeling performance for the generated explanations in the interpretation layer but also excels in producing efficient and effective prediction sets for the final predictions based on the informative high-level basis explanations. We perform the theoretical analysis for the proposed framework. Extensive experimental evaluation demonstrates the effectiveness of the proposed uncertainty framework.

re-training (MEDCSP), a new pre-training strategy designed to bridge the gap between multimodal medical sources. MEDCSP employs modality-level aggregation to unify patient data within individual sources. Additionally, leveraging temporal information and diagnosis history, MEDCSP effectively captures explicit and implicit correlations between patients across different sources. To evaluate the proposed strategy, we conduct comprehensive experiments, where the experiments are based on 6 modalities from 2 real-world medical data sources, and MEDCSP is evaluated on 4 tasks against 19 baselines, marking an initial yet essential step towards cross-source modeling in the medical domain.

Propelled by their remarkable capabilities to generate novel and engaging content, Generative Artificial Intelligence (GenAI) technologies are disrupting traditional workflows in many industries. While prior research has examined GenAI from a techno-centric perspective, there is still a lack of understanding about how users perceive and utilize GenAI in real-world scenarios. To bridge this gap, we conducted semi-structured interviews with (N = 18) GenAI users in creative industries, investigating the human-GenAI co-creation process within a holistic LUA (Learning, Using and Assessing) framework. Our study uncovered an intriguingly complex landscape: Prospects – GenAI greatly fosters the co-creation between human expertise and GenAI capabilities, profoundly transforming creative workflows; Challenges – Meanwhile, users face substantial uncertainties and complexities arising from resource availability, tool usability, and regulatory compliance; Strategies – In response, users actively devise various strategies to overcome many of such challenges. Our study reveals key implications for the design of future GenAI tools.

Automatic International Classification of Diseases (ICD) coding plays a crucial role in the extraction of relevant information from clinical notes for proper recording and billing. One of the most important directions for boosting the performance of automatic ICD coding is modeling ICD code relations. However, current methods insufficiently model the intricate relationships among ICD codes and often overlook the importance of context in clinical notes. In this paper, we propose a novel approach, a contextualized and flexible framework, to enhance the learning of ICD code representations. Our approach, unlike existing methods, employs a dependent learning paradigm that considers the context of clinical notes in modeling all possible code relations. We evaluate our approach on six public ICD coding datasets and the experimental results demonstrate the effectiveness of our approach compared to state-of-the-art baselines.

Foundation models have demonstrated remarkable capabilities in handling diverse modalities and tasks, outperforming conventional artificial intelligence (AI) approaches that are highly task-specific and modality-reliant.In the medical domain, however, the development of comprehensive foundation models is constrained by limited access to diverse modalities and stringent privacy regulations.To address these constraints, this study introduces a novel knowledge injection approach, FED-KIM, designed to scale the medical foundation model within a federated learning framework.FEDKIM leverages lightweight local models to extract healthcare knowledge from private data and integrates this knowledge into a centralized foundation model using a designed adaptive Multitask Multimodal Mixture Of Experts (M 3 OE) module.This method not only preserves privacy but also enhances the model's ability to handle complex medical tasks involving multiple modalities.Our extensive experiments across twelve tasks in seven modalities demonstrate the effectiveness of FEDKIM in various settings, highlighting its potential to scale medical foundation models without direct access to sensitive data.

This paper presents FedType, a simple yet pioneering framework designed to fill research gaps in heterogeneous model aggregation within federated learning (FL). FedType introduces small identical proxy models for clients, serving as agents for information exchange, ensuring model security, and achieving efficient communication simultaneously. To transfer knowledge between large private and small proxy models on clients, we propose a novel uncertainty-based asymmetrical reciprocity learning method, eliminating the need for any public data. Comprehensive experiments conducted on benchmark datasets demonstrate the efficacy and generalization ability of FedType across diverse settings. Our approach redefines federated learning paradigms by bridging model heterogeneity, eliminating reliance on public data, prioritizing client privacy, and reducing communication costs (The codes are available in the supplementation materials).

Efficient and accurate product relevance assessment is critical for user experiences and business success. Training a proficient relevance assessment model requires high-quality query-product pairs, often obtained through negative sampling strategies. Unfortunately, current methods introduce pooling bias by mistakenly sampling false negatives, diminishing performance and business impact. To address this, we present Bias-mitigating Hard Negative Sampling (BHNS), a novel negative sampling strategy tailored to identify and adjust for false negatives, building upon our original False Negative Estimation algorithm. Our experiments in the Instacart search setting confirm BHNS as effective for practical e-commerce use. Furthermore, comparative analyses on public dataset showcase its domain-agnostic potential for diverse applications.

Synthesizing electronic health records (EHR) data has become a preferred strategy to address data scarcity, improve data quality, and model fairness in healthcare. However, existing approaches for EHR data generation predominantly rely on state-of-the-art generative techniques like generative adversarial networks, variational autoencoders, and language models. These methods typically replicate input visits, resulting in inadequate modeling of temporal dependencies between visits and overlooking the generation of time information, a crucial element in EHR data. Moreover, their ability to learn visit representations is limited due to simple linear mapping functions, thus compromising generation quality. To address these limitations, we propose a novel EHR data generation model called EHRPD. It is a diffusion-based model designed to predict the next visit based on the current one while also incorporating time interval estimation. To enhance generation quality and diversity, we introduce a novel time-aware visit embedding module and a pioneering predictive denoising diffusion probabilistic model (P-DDPM). Additionally, we devise a predictive U-Net (PU-Net) to optimize P-DDPM. We conduct experiments on two public datasets and evaluate EHRPD from fidelity, privacy, and utility perspectives. The experimental results demonstrate the efficacy and utility of the proposed EHRPD in addressing the aforementioned limitations and advancing EHR data generation.

This study introduces the Federated Medical Knowledge Injection (FedMEKI) platform, a new benchmark designed to address the unique challenges of integrating medical knowledge into foundation models under privacy constraints. By leveraging a cross-silo federated learning approach, FedMEKI circumvents the issues associated with centralized data collection, which is often prohibited under health regulations like the Health Insurance Portability and Accountability Act (HIPAA) in the USA. The platform is meticulously designed to handle multi-site, multi-modal, and multi-task medical data, which includes 7 medical modalities, including images, signals, texts, laboratory test results, vital signs, input variables, and output variables. The curated dataset to validate FedMEKI covers 8 medical tasks, including 6 classification tasks (lung opacity detection, COVID-19 detection, electrocardiogram (ECG) abnormal detection, mortality prediction, sepsis protection, and enlarged cardiomediastinum detection) and 2 generation tasks (medical visual question answering (MedVQA) and ECG noise clarification). This comprehensive dataset is partitioned across several clients to facilitate the decentralized training process under 16 benchmark approaches. FedMEKI not only preserves data privacy but also enhances the capability of medical foundation models by allowing them to learn from a broader spectrum of medical knowledge without direct data exposure, thereby setting a new benchmark in the application of foundation models within the healthcare sector.

Researchers have been studying approaches to steer the behavior of Large Language Models (LLMs) and build personalized LLMs tailored for various applications. While fine-tuning seems to be a direct solution, it requires substantial computational resources and may significantly affect the utility of the original LLM. Recent endeavors have introduced more lightweight strategies, focusing on extracting ``steering vectors'' to guide the model's output toward desired behaviors by adjusting activations within specific layers of the LLM's transformer architecture. However, such steering vectors are directly extracted from the activations of human preference data and thus often lead to suboptimal results and occasional failures, especially in alignment-related scenarios. In this work, we propose an innovative approach that could produce more effective steering vectors through bi-directional preference optimization. Our method is designed to allow steering vectors to directly influence the generation probability of contrastive human preference data pairs, thereby offering a more precise representation of the target behavior. By carefully adjusting the direction and magnitude of the steering vector, we enabled personalized control over the desired behavior across a spectrum of intensities. Extensive experimentation across various open-ended generation tasks, particularly focusing on steering AI personas, has validated the efficacy of our approach. Moreover, we comprehensively investigate critical alignment-concerning scenarios, such as managing truthfulness, mitigating hallucination, and addressing jailbreaking attacks alongside their respective defenses. Remarkably, our method can still demonstrate outstanding steering effectiveness across these scenarios. Furthermore, we showcase the transferability of our steering vectors across different models/LoRAs and highlight the synergistic benefits of applying multiple vectors simultaneously. These findings significantly broaden the practicality and versatility of our proposed method.

Federated learning, a pioneering paradigm, enables collaborative model training without exposing users’ data to central servers. Most existing federated learning systems necessitate uniform model structures across all clients, restricting their practicality. Several methods have emerged to aggregate diverse client models; however, they either lack the ability of personalization, raise privacy and security concerns, need prior knowledge, or ignore the capability and functionality of personalized models. In this paper, we present an innovative approach, named pFedClub, which addresses these challenges. pFedClub introduces personalized federated learning through the substitution of controllable neural network blocks/layers. Initially, pFedClub dissects heterogeneous client models into blocks and organizes them into functional groups on the server. Utilizing the designed CMSR (Controllable Model Searching and Reproduction) algorithm, pFedClub generates a range of personalized candidate models for each client. A model matching technique is then applied to select the optimal personalized model, serving as a teacher model to guide each client’s training process. We conducted extensive experiments across three datasets, examining both IID and non-IID settings. The results demonstrate that pFedClub outperforms baseline approaches, achieving state-of-the-art performance. Moreover, our model insight analysis reveals that pFedClub generates personalized models of reasonable size in a controllable manner, significantly reducing computational costs.

Sleep apnea is a serious sleep disorder where patients have multiple extended pauses in breath during sleep. Although some portable or contactless sleep apnea detection systems have been proposed, none of them can achieve fine-grained sleep apnea detection without strict requirements on the device or environmental settings. To address this problem, we present DeepApnea, a deep learning based sleep apnea detection system that leverages patients' wrist movement data collected by smartwatches to identify different types of sleep apnea events (i.e., central apneas, obstructive apneas, and hypopneas). Through a clinical study, we identify some special characteristics associated with different types of sleep apnea captured by smartwatch. However, there are many technical challenges such as how to extract informative apnea features from the noisy data and how to leverage features extracted from the multi-axis sensing data. To address these challenges, we first propose signal pre-processing methods to filter the raw accelerometer (ACC) data, smoothing away noise while preserving the respiratory signal and potential features for identifying sleep apnea. Then, we design a deep learning architecture to extract features from three ACC axes collaboratively, where self attention and cross-axis correlation techniques are leveraged to improve the classification accuracy. We have implemented DeepApnea on smartwatches and performed a clinical study. Evaluation results demonstrate that DeepApnea can significantly outperform existing work on identifying different types of sleep apnea.

The widespread adoption of Electronic Health Record (EHR) systems in healthcare institutes has generated vast amounts of medical data, offering significant opportunities for improving healthcare services through deep learning techniques. However, the complex and diverse modalities and feature structures in real-world EHR data pose great challenges for deep learning model design. To address the multi-modality challenge in EHR data, current approaches primarily rely on hand-crafted model architectures based on intuition and empirical experiences, leading to sub-optimal model architectures and limited performance. Therefore, to automate the process of model design for mining EHR data, we propose a novel neural architecture search (NAS) framework named AutoFM, which can automatically search for the optimal model architectures for encoding diverse input modalities and fusion strategies. We conduct thorough experiments on real-world multi-modal EHR data and prediction tasks, and the results demonstrate that our framework not only achieves significant performance improvement over existing state-of-the-art methods but also discovers meaningful network architectures effectively.

Health risk prediction aims to forecast the potential health risks that patients may face using their historical Electronic Health Records (EHR). Although several effective models have developed, data insufficiency is a key issue undermining their effectiveness. Various data generation and augmentation methods have been introduced to mitigate this issue by expanding the size of the training data set through learning underlying data distributions. However, the performance of these methods is often limited due to their task-unrelated design. To address these shortcomings, this paper introduces a novel, end-to-end diffusion-based risk prediction model, named MedDiffusion. It enhances risk prediction performance by creating synthetic patient data during training to enlarge sample space. Furthermore, MedDiffusion discerns hidden relationships between patient visits using a step-wise attention mechanism, enabling the model to automatically retain the most vital information for generating high-quality data. Experimental evaluation on four real-world medical datasets demonstrates that MedDiffusion outperforms 14 cutting-edge baselines in terms of PR-AUC, F1, and Cohen's Kappa. We also conduct ablation studies and benchmark our model against GAN-based alternatives to further validate the rationality and adaptability of our model design. Additionally, we analyze generated data to offer fresh insights into the model's interpretability. The source code is available via https://shorturl.at/aerT0.

Clinical trial analysis is one of the main business directions and services in IQVIA, and reviewing past similar studies is one of the most critical steps before starting a commercial clinical trial. The current review process is manual and time-consuming, requiring a clinical trial analyst to manually search through an extensive clinical trial database and then review all candidate studies. Therefore, it is of great interest to develop an automatic retrieval algorithm to select similar studies by giving new study information. To achieve this goal, we propose a novel group-based trial similarity learning network named GTSLNet, consisting of two kinds of similarity learning modules. The pair-wise section-level similarity learning module aims to compare the query trial and the candidate trial from the abstract semantic level via the proposed section transformer. Meanwhile, a word-level similarity learning module uses the word similarly matrix to capture the low-level similarity information. Additionally, an aggregation module combines these similarities. To address potential false negatives and noisy data, we introduce a variance-regularized group distance loss function. Experiment results show that the proposed GTSLNet significantly and consistently outperforms state-of-the-art baselines.

Session-based recommendation aims to predict intents of anonymous users based on their limited behaviors. Modeling user behaviors involves two distinct rationales: co-occurrence patterns reflected by item IDs, and fine-grained preferences represented by item modalities (e.g., text and images). However, existing methods typically entangle these causes, leading to their failure in achieving accurate and explainable recommendations. To this end, we propose a novel framework DIMO to disentangle the effects of ID and modality in the task. DIMO aims to disentangle these causes at both item and session levels. At the item level, we introduce a co-occurrence representation schema to explicitly incorporate co-occurrence patterns into ID representations. Simultaneously, DIMO aligns different modalities into a unified semantic space to represent them uniformly. At the session level, we present a multi-view self-supervised disentanglement, including proxy mechanism and counterfactual inference, to disentangle ID and modality effects without supervised signals. Leveraging these disentangled causes, DIMO provides recommendations via causal inference and further creates two templates for generating explanations. Extensive experiments on multiple real-world datasets demonstrate the consistent superiority of DIMO over existing methods. Further analysis also confirms DIMO's effectiveness in generating explanations.

Sequential recommendation is dedicated to offering items of interest for users based on their history behaviors. The attribute-opinion pairs, expressed by users in their reviews for items, provide the potentials to capture user preferences and item characteristics at a fine-grained level. To this end, we propose a novel framework FineRec that explores the attribute-opinion pairs of reviews to finely handle sequential recommendation. Specifically, we utilize a large language model to extract attribute-opinion pairs from reviews. For each attribute, a unique attribute-specific user-opinion-item graph is created, where corresponding opinions serve as the edges linking heterogeneous user and item nodes. Afterwards, we devise a diversity-aware convolution operation to aggregate information within the graphs, enabling attribute-specific user and item representation learning. Ultimately, we present an interaction-driven fusion mechanism to integrate attribute-specific user/item representations across all attributes for generating recommendations. Extensive experiments conducted on several real-world datasets demonstrate the superiority of our FineRec over existing state-ofthe-art methods. Further analysis also verifies the effectiveness of our fine-grained manner in handling the task.

Session-based recommendation is devoted to characterizing preferences of anonymous users based on short sessions. Existing methods mostly focus on mining limited item co-occurrence patterns exposed by item ID within sessions, while ignoring what attracts users to engage with certain items is rich multi-modal information displayed on pages. Generally, the multi-modal information can be classified into two categories: descriptive information (e.g., item images and description text) and numerical information (e.g., price). In this paper, we aim to improve session-based recommendation by modeling the above multi-modal information holistically. There are mainly three issues to reveal user intent from multi-modal information: (1) How to extract relevant semantics from heterogeneous descriptive information with different noise? (2) How to fuse these heterogeneous descriptive information to comprehensively infer user interests? (3) How to handle probabilistic influence of numerical information on user behaviors? To solve above issues, we propose a novel multi-modal session-based recommendation (MMSBR) that models both descriptive and numerical information under a unified framework. Specifically, a pseudo-modality contrastive learning is devised to enhance the representation learning of descriptive information. Afterwards, a hierarchical pivot transformer is presented to fuse heterogeneous descriptive information. Moreover, we represent numerical information with Gaussian distribution and design a Wasserstein self-attention to handle the probabilistic influence mode. Extensive experiments on three real-world datasets demonstrate the effectiveness of the proposed MMSBR. Further analysis also proves that our MMSBR can alleviate the cold-start problem in SBR effectively.

Session-based recommendation intends to predict next purchased items based on anonymous behavior sequences. Numerous economic studies have revealed that item price is a key factor influencing user purchase decisions. Unfortunately, existing methods for session-based recommendation only aim at capturing user interest preference, while ignoring user price preference. Actually, there are primarily two challenges preventing us from accessing price preference. First, the price preference is highly associated to various item features (i.e., category and brand), which asks us to mine price preference from heterogeneous information. Second, price preference and interest preference are interdependent and collectively determine user choice, necessitating that we jointly consider both price and interest preference for intent modeling. To handle above challenges, we propose a novel approach Bi-Preference Learning Heterogeneous Hypergraph Networks (BiPNet) for session-based recommendation. Specifically, the customized heterogeneous hypergraph networks with a triple-level convolution are devised to capture user price and interest preference from heterogeneous features of items. Besides, we develop a Bi-Preference Learning schema to explore mutual relations between price and interest preference and collectively learn these two preferences under the multi-task learning architecture. Extensive experiments on multiple public datasets confirm the superiority of BiPNet over competitive baselines. Additional research also supports the notion that the price is crucial for the task.

Hard-label textual adversarial attack is a challenging task, as only the predicted label information is available, and the text space is discrete and non-differentiable. Relevant research work is still in fancy and just a handful of methods are proposed. However, existing methods suffer from either the high complexity of genetic algorithms or inaccurate gradient estimation, thus are arduous to obtain adversarial examples with high semantic similarity and low perturbation rate under the tight-budget scenario. In this paper, we propose a simple and sweet paradigm for hard-label textual adversarial attack, named SSPAttack. Specifically, SSPAttack first utilizes initialization to generate an adversarial example, and removes unnecessary replacement words to reduce the number of changed words. Then it determines the replacement order and searches for an anchor synonym, thus avoiding going through all the synonyms. Finally, it pushes substitution words towards original words until an appropriate adversarial example is obtained. The core idea of SSPAttack is just swapping words whose mechanism is simple. Experimental results on eight benchmark datasets and two real-world APIs have shown that the performance of SSPAttack is sweet in terms of similarity, perturbation rate and query efficiency.

Distribution estimation has been demonstrated as one of the most effective approaches in dealing with few-shot image classification, as the low-level patterns and underlying representations can be easily transferred across different tasks in computer vision domain. However, directly applying this approach to few-shot text classification is challenging, since leveraging the statistics of known classes with sufficient samples to calibrate the distributions of novel classes may cause negative effects due to serious category difference in text domain. To alleviate this issue, we propose two simple yet effective strategies to estimate the distributions of the novel classes by utilizing unlabeled query samples, thus avoiding the potential negative transfer issue. Specifically, we first assume a class or sample follows the Gaussian distribution, and use the original support set and the nearest few query samples to estimate the corresponding mean and covariance. Then, we augment the labeled samples by sampling from the estimated distribution, which can provide sufficient supervision for training the classification model. Extensive experiments on eight few-shot text classification datasets show that the proposed method outperforms state-of-the-art baselines significantly.

Automated machine learning (AutoML) is envisioned to make ML techniques accessible to ordinary users. Recent work has investigated the role of humans in enhancing AutoML functionality throughout a standard ML workflow. However, it is also critical to understand how users adopt existing AutoML solutions in complex, real-world settings from a holistic perspective. To fill this gap, this study conducted semi-structured interviews of AutoML users (N = 19) focusing on understanding (1) the limitations of AutoML encountered by users in their real-world practices, (2) the strategies users adopt to cope with such limitations, and (3) how the limitations and workarounds impact their use of AutoML. Our findings reveal that users actively exercise user agency to overcome three major challenges arising from customizability, transparency, and privacy. Furthermore, users make cautious decisions about whether and how to apply AutoML on a case-by-case basis. Finally, we derive design implications for developing future AutoML solutions.

Meta-learning methods have shown to be effective in dealing with cold-start recommendation. However, most previous methods rely on an ideal assumption that there exists a similar data distribution between source and target tasks, which are unsuitable for the scenario that only extremely limited number of new user or item interactions are available. In this paper, we propose to boost meta-learning cold-start recommendation with graph neural network (MeGNN). First, it utilizes the global neighborhood translation learning to obtain consistent potential interactions for all new user and item nodes, which can refine their representations. Second, it employs the local neighborhood translation learning to predict specific potential interactions for each node, thus guaranteeing the personalized requirement. In experiments, we combine MeGNN with two representative meta-learning models MeLU and TaNP. Extensive results on two widely-used datasets show the superiority of MeGNN in four different scenarios.

Clinical trials aim to study new tests and evaluate their effects on human health outcomes, which has a huge market size. However, carrying out clinical trials is expensive and time-consuming and often ends in no results. It will revolutionize clinical practice if we can develop an effective model to automatically estimate the status of a clinical trial and find out possible failure reasons. However, it is challenging to develop such a model because of the lack of a benchmark dataset. To address these challenges, in this paper, we first build a new dataset by extracting the publicly available clinical trial reports from ClinicalTrials.gov. The associated status of each report is treated as the status label. To analyze the failure reasons, domain experts help us manually annotate each failed report based on the description associated with it. More importantly, we examine several state-of-the-art text classification baselines on this task and find out that the unique format of the clinical trial protocols plays an essential role in affecting prediction accuracy, demonstrating the need for specially designed clinical trial classification models.

Predicting adverse drug reactions (ADRs) of drugs is one of the most critical steps in drug development. By pre-estimating the adverse reactions, researchers and drug development companies can greatly prevent the potential ADR risks and tragedies. However, the current ADR prediction methods suffer from several limitations. First, the prediction results are based on pure drug-related information, which makes them impossible to be directly applied for the personalized ADR prediction task. The lack of personalization of models also makes rare adverse events hard to be predicted. Therefore, it is of great interest to develop a new personalized ADR prediction method by introducing additional sources, e.g., patient health records. However, few methods have tried to use additional sources. In the meantime, the variety of different source formats and structures makes this task more challenging. To address the above challenges, we propose a novel personalized multi-sourced-based drug adverse reaction prediction model named pADR. pADR first works on every single source to transform them into proper representations. Next, a hierarchical multi-sourced Transformer is designed to automatically model the interactions between different sources and fuse them together for the final adverse event prediction. Experimental results on a new multi-sourced ADR prediction dataset show that PADR outperforms state-of-the-art drug-based baselines. Moreover, the case and ablation studies also illustrate the effectiveness of our proposed fusion strategies and the reasonableness of each module design.

Pretraining has proven to be a powerful technique in natural language processing (NLP), exhibiting remarkable success in various NLP downstream tasks. However, in the medical domain, existing pretrained models on electronic health records (EHR) fail to capture the hierarchical nature of EHR data, limiting their generalization capability across diverse downstream tasks using a single pretrained model. To tackle this challenge, this paper introduces a novel, general, and unified pretraining framework called MedHMP, specifically designed for hierarchically multimodal EHR data. The effectiveness of the proposed MedHMP is demonstrated through experimental results on eight downstream tasks spanning three levels. Comparisons against eighteen baselines further highlight the efficacy of our approach.

Decision-based methods have shown to be effective in black-box adversarial attacks, as they can obtain satisfactory performance and only require to access the final model prediction. Gradient estimation is a critical step in black-box adversarial attacks, as it will directly affect the query efficiency. Recent works have attempted to utilize gradient priors to facilitate score-based methods to obtain better results. However, these gradient priors still suffer from the edge gradient discrepancy issue and the successive iteration gradient direction issue, thus are difficult to simply extend to decision-based methods. In this paper, we propose a novel Decision-based Black-box Attack framework with Gradient Priors (DBA-GP), which seamlessly integrates the data-dependent gradient prior and time-dependent prior into the gradient estimation procedure. First, by leveraging the joint bilateral filter to deal with each random perturbation, DBA-GP can guarantee that the generated perturbations in edge locations are hardly smoothed, i.e., alleviating the edge gradient discrepancy, thus remaining the characteristics of the original image as much as possible. Second, by utilizing a new gradient updating strategy to automatically adjust the successive iteration gradient direction, DBA-GP can accelerate the convergence speed, thus improving the query efficiency. Extensive experiments have demonstrated that the proposed method outperforms other strong baselines significantly.

Despite a plethora of prior explorations, conducting text adversarial attacks in practical settings is still challenging with the following constraints: black box -- the inner structure of the victim model is unknown; hard label -- the attacker only has access to the top-1 prediction results; and semantic preservation - the perturbation needs to preserve the original semantics. In this paper, we present PAT, a novel adversarial attack method employed under all these constraints. Specifically, PAT explicitly models the adversarial and non-adversarial prototypes and incorporates them to measure semantic changes for replacement selection in the hard-label black-box setting to generate high-quality samples. In each iteration, PAT finds original words that can be replaced back and selects better candidate words for perturbed positions in a geometry-aware manner guided by this estimation, which maximally improves the perturbation construction and minimally impacts the original semantics. Extensive evaluation with benchmark datasets and state-of-the-art models shows that PAT outperforms existing text adversarial attacks in terms of both attack effectiveness and semantic preservation. Moreover, we validate the efficacy of PAT against industry-leading natural language processing platforms in real-world settings.

Vision-Language (VL) pre-trained models have shown their superiority on many multimodal tasks. However, the adversarial robustness of such models has not been fully explored. Existing approaches mainly focus on exploring the adversarial robustness under the white-box setting, which is unrealistic. In this paper, we aim to investigate a new yet practical task to craft image and text perturbations using pre-trained VL models to attack black-box fine-tuned models on different downstream tasks. Towards this end, we propose VLATTACK to generate adversarial samples by fusing perturbations of images and texts from both single-modal and multi-modal levels. At the single-modal level, we propose a new block-wise similarity attack (BSA) strategy to learn image perturbations for disrupting universal representations. Besides, we adopt an existing text attack strategy to generate text perturbations independent of the image-modal attack. At the multi-modal level, we design a novel iterative cross-search attack (ICSA) method to update adversarial image-text pairs periodically, starting with the outputs from the single-modal level. We conduct extensive experiments to attack three widely-used VL pretrained models for six tasks on eight datasets. Experimental results show that the proposed VLATTACK framework achieves the highest attack success rates on all tasks compared with state-of-the-art baselines, which reveals a significant blind spot in the deployment of pre-trained VL models.

Black-box hard-label adversarial attack on text is a practical and challenging task, as the text data space is inherently discrete and non-differentiable, and only the predicted label is accessible. Research on this problem is still in the embryonic stage and only a few methods are available. Nevertheless, existing methods rely on the complex heuristic algorithm or unreliable gradient estimation strategy, which probably fall into the local optimum and inevitably consume numerous queries, thus are difficult to craft satisfactory adversarial examples with high semantic similarity and low perturbation rate in a limited query budget. To alleviate above issues, we propose a simple yet effective framework to generate high quality textual adversarial examples under the black-box hard-label attack scenarios, named HQA-Attack. Specifically, after initializing an adversarial example randomly, HQA-attack first constantly substitutes original words back as many as possible, thus shrinking the perturbation rate. Then it leverages the synonym set of the remaining changed words to further optimize the adversarial example with the direction which can improve the semantic similarity and satisfy the adversarial condition simultaneously. In addition, during the optimizing procedure, it searches a transition synonym word for each changed word, thus avoiding traversing the whole synonym set and reducing the query number to some extent. Extensive experimental results on five text classification datasets, three natural language inference datasets and two real-world APIs have shown that the proposed HQA-Attack method outperforms other strong baselines significantly.

This paper focuses on addressing the practical yet challenging problem of model heterogeneity in federated learning, where clients possess models with different network structures. To track this problem, we propose a novel framework called pFedHR, which leverages heterogeneous model reassembly to achieve personalized federated learning. In particular, we approach the problem of heterogeneous model personalization as a model-matching optimization task on the server side. Moreover, pFedHR automatically and dynamically generates informative and diverse personalized candidates with minimal human intervention. Furthermore, our proposed heterogeneous model reassembly technique mitigates the adverse impact introduced by using public data with different distributions from the client data to a certain extent. Experimental results demonstrate that pFedHR outperforms baselines on three datasets under both IID and Non-IID settings. Additionally, pFedHR effectively reduces the adverse impact of using different public data and dynamically generates diverse personalized models in an automated manner.

Pre-trained language models (PLMs) have demonstrated remarkable performance as few-shot learners. However, their security risks under such settings are largely unexplored. In this work, we conduct a pilot study showing that PLMs as few-shot learners are highly vulnerable to backdoor attacks while existing defenses are inadequate due to the unique challenges of few-shot scenarios. To address such challenges, we advocate MDP, a novel lightweight, pluggable, and effective defense for PLMs as few-shot learners. Specifically, MDP leverages the gap between the masking-sensitivity of poisoned and clean samples: with reference to the limited few-shot data as distributional anchors, it compares the representations of given samples under varying masking and identifies poisoned samples as ones with significant variations. We show analytically that MDP creates an interesting dilemma for the attacker to choose between attack effectiveness and detection evasiveness. The empirical evaluation using benchmark datasets and representative attacks validates the efficacy of MDP. The code of MDP is publicly available.

Recent years have witnessed a surge of certified robust training pipelines against text adversarial perturbation constructed by synonym substitutions. Given a base model, existing pipelines provide prediction certificates either in the discrete word space or the continuous latent space. However, they are isolated from each other with a structural gap. We observe that existing training frameworks need unification to provide stronger certified robustness. Additionally, they mainly focus on building the certification process but neglect to improve the robustness of the base model. To mitigate the aforementioned limitations, we propose a unified framework named UniT that enables us to train flexibly in either fashion by working in the word embedding space. It can provide a stronger robustness guarantee obtained directly from the word embedding space without extra modules. In addition, we introduce the decoupled regularization (DR) loss to improve the robustness of the base model, which includes two separate robustness regularization terms for the feature extraction and classifier modules. Experimental results on widely used text classification datasets further demonstrate the effectiveness of the designed unified framework and the proposed DR loss for improving the certified robust accuracy.

Federated learning (FL) enables multiple clients to train models collaboratively without sharing local data, which has achieved promising results in different areas, including the Internet of Things (IoT). However, end IoT devices do not have abilities to automatically annotate their collected data, which leads to the label shortage issue at the client side. To collaboratively train an FL model, we can only use a small number of labeled data stored on the server. This is a new yet practical scenario in federated learning, i.e., labels-at-server semi-supervised federated learning (SemiFL). Although several SemiFL approaches have been proposed recently, none of them can focus on the personalization issue in their model design. IoT environments make SemiFL more challenging, as we need to take device computational constraints and communication cost into consideration simultaneously. To tackle these new challenges together, we propose a novel SemiFL framework named pFedKnow. pFedKnow generates lightweight personalized client models via neural network pruning techniques to reduce communication cost. Moreover, it incorporates pretrained large models as prior knowledge to guide the aggregation of personalized client models and further enhance the framework performance. Experiment results on both image and text datasets show that the proposed pFedKnow outperforms state-of-the-art baselines as well as reducing considerable communication cost. The source code of the proposed pFedKnow is available at https://github.com/JackqqWang/pfedknow/tree/master.

The fast spread of hate speech on social media impacts the Internet environment and our society by increasing prejudice and hurting people. Detecting hate speech has aroused broad attention in the field of natural language processing. Although hate speech detection has been addressed in recent work, this task still faces two inherent unsolved challenges. The first challenge lies in the complex semantic information conveyed in hate speech, particularly the interference of insulting words in hate speech detection. The second challenge is the imbalanced distribution of hate speech and non-hate speech, which may significantly deteriorate the performance of models. To tackle these challenges, we propose a novel dual contrastive learning (DCL) framework for hate speech detection. Our framework jointly optimizes the self-supervised and the supervised contrastive learning loss for capturing span-level information beyond the token-level emotional semantics used in existing models, particularly detecting speech containing abusive and insulting words. Moreover, we integrate the focal loss into the dual contrastive learning framework to alleviate the problem of data imbalance. We conduct experiments on two publicly available English datasets, and experimental results show that the proposed model outperforms the state-of-the-art models and precisely detects hate speeches.

The increasing reliance on online communities for healthcare information by patients and caregivers has led to the increase in the spread of misinformation, or subjective, anecdotal and inaccurate or non-specific recommendations, which, if acted on, could cause serious harm to the patients. Hence, there is an urgent need to connect users with accurate and tailored health information in a timely manner to prevent such harm. This article proposes an innovative approach to suggesting reliable information to participants in online communities as they move through different stages in their disease or treatment. We hypothesize that patients with similar histories of disease progression or course of treatment would have similar information needs at comparable stages. Specifically, we pose the problem of predicting topic tags or keywords that describe the future information needs of users based on their profiles, traces of their online interactions within the community (past posts, replies) and the profiles and traces of online interactions of other users with similar profiles and similar traces of past interaction with the target users. The result is a variant of the collaborative information filtering or recommendation system tailored to the needs of users of online health communities. We report results of our experiments on two unique datasets from two different social media platforms which demonstrates the superiority of the proposed approach over the state of the art baselines with respect to accurate and timely prediction of topic tags (and hence information sources of interest).

Chatbot-based symptom checker (CSC) apps have become increasingly popular in healthcare. These apps engage users in human-like conversations and offer possible medical diagnoses. The conversational design of these apps can significantly impact user perceptions and experiences, and may influence medical decisions users make and the medical care they receive. However, the effects of the conversational design of CSCs remain understudied, and there is a need to investigate and enhance users’ interactions with CSCs. In this article, we conducted a two-stage exploratory study using a human-centered design methodology. We first conducted a qualitative interview study to identify key user needs in engaging with CSCs. We then performed an experimental study to investigate potential CSC conversational design solutions based on the results from the interview study. We identified that emotional support, explanations of medical information, and efficiency were important factors for users in their interactions with CSCs. We also demonstrated that emotional support and explanations could affect user perceptions and experiences, and they are context-dependent. Based on these findings, we offer design implications for CSC conversations to improve the user experience and health-related decision-making.

This paper focuses on a newly challenging setting in hard-label adversarial attacks on text data by taking the budget information into account. Although existing approaches can successfully generate adversarial examples in the hard-label setting, they follow an ideal assumption that the victim model does not restrict the number of queries. However, in real-world applications the query budget is usually tight or limited. Moreover, existing hard-label adversarial attack techniques use the genetic algorithm to optimize discrete text data by maintaining a number of adversarial candidates during optimization, which can lead to the problem of generating low-quality adversarial examples in the tight-budget setting. To solve this problem, in this paper, we propose a new method named TextHoaxer by formulating the budgeted hard-label adversarial attack task on text data as a gradient-based optimization problem of perturbation matrix in the continuous word embedding space. Compared with the genetic algorithm-based optimization, our solution only uses a single initialized adversarial example as the adversarial candidate for optimization, which significantly reduces the number of queries. The optimization is guided by a new objective function consisting of three terms, i.e., semantic similarity term, pair-wise perturbation constraint, and sparsity constraint. Semantic similarity term and pair-wise perturbation constraint can ensure the high semantic similarity of adversarial examples from both comprehensive text-level and individual word-level, while the sparsity constraint explicitly restricts the number of perturbed words, which is also helpful for enhancing the quality of generated text. We conduct extensive experiments on eight text datasets against three representative natural language models, and experimental results show that TextHoaxer can generate high-quality adversarial examples with higher semantic similarity and lower perturbation rate under the tight-budget setting.

Electronic health records (EHR) have been widely applied to various tasks in the medical domain such as risk predictive modeling, which aims to predict further health conditions by analyzing patients' historical EHR. Existing work mainly focuses on modeling the sequential and temporal characteristics of EHR data with advanced deep learning techniques. However, the network architectures of these models are all manually designed based on experts' prior knowledge, which largely impedes non-experts from exploring this task. To address this issue, in this paper, we propose a novel automated risk prediction model named AutoMed to automatically search the optimal model architecture for modeling the complex EHR data and improving the performance of the risk prediction task. In particular, we follow the idea of neural architecture search to design a search space that contains three separate searchable modules. Two of them are used for analyzing sequential and temporal features of EHR data, respectively. The third is to automatically fuse both features together. Besides these three modules, AutoMed contains an embedding module and a prediction module. All the three searchable modules are jointly optimized in the search stage to derive the optimal model architecture. In such a way, the model design can be automatically achieved with few human interventions. Experimental results on three real-world datasets show that AutoMed outperforms state-of-the-art baselines in terms of PR-AUC, F1, and Cohen's Kappa. Moreover, the ablation study shows that AutoMed can obtain reasonable model architectures and offer useful insights to the future risk prediction model design.

Numerous people are suffering from sleep-related problems. To diagnose them, a prerequisite is to divide the polysomnography (PSG) data into different sleep stages. Thus, sleep stage classification is an essential step, but collecting PSG data is expensive, time-consuming, and even belated. To address this issue, using accelerometers that are widely used in smartwatches is treated as an alternative way to monitor people’s sleep conditions. However, the flexibility of deep learning models by purely using wrist-worn accelerometer data for sleep stage classification has not been investigated by researchers. To explore the answer, in this paper, we design a novel axis-aware hybrid fusion-based deep learning model, named AccSleepNet, which takes the three axes’ accelerometer data as the input simultaneously. The designed axis-aware hybrid fusion mechanism prompts the model to learn the deep features from three axes collaboratively. Finally, a classification module takes the fused feature representations from three axes as input and outputs the predicted sleep stage. Experimental results on two public datasets demonstrate the effectiveness of the proposed AccSleepNet for the sleep stage classification task compared with state-of-the-art baselines. Moreover, an ablation study validates the necessity of leveraging three axes’ accelerometer data and the superiority of the designed axis-aware hybrid fusion mechanism1.

Researchers have conduct adversarial attacks against deep neural networks (DNNs) for health risk prediction in the white/gray-box setting to evaluate their robustness. However, since most real-world solutions are trained by private data and released as black-box services on the cloud, we should investigate their robustness in the black-box setting. Unfortunately, existing work ignores to consider the uniqueness of electronic health records (EHRs). To fill this gap, we propose the first black-box adversarial attack method against health risk prediction models named MedAttacker to investigate their vulnerability. It addresses the challenges brought by EHRs via two steps: hierarchical position selection which selects the attacked positions in a reinforcement learning (RL) framework and substitute selection which identifies substitutes with a score-based principle. Particularly, by considering the temporal context inside EHRs, MedAttacker initializes its RL position selection policy by using the contribution score of each visit and the saliency score of each code, which can be well integrated with the deterministic substitute selection process decided by the score changes. We evaluate MedAttacker by attacking three advanced risk prediction models in the black-box setting across multiple real-world datasets, and MedAttacker consistently achieves the highest average success rate and even outperforms a recent white-box EHR adversarial attack technique in certain cases.

Intelligent medical services have attracted great research interests for providing automated medical consultation. However, the lack of corpora becomes a main obstacle to related research, particularly data from real scenarios. In this paper, we construct RealMedDial, a Chinese medical dialogue dataset based on real medical consultation. RealMedDial contains 2,637 medical dialogues and 24,255 utterances obtained from Chinese short-video clips of real medical consultations. We collected and annotated a wide range of meta-data with respect to medical dialogue including doctor profiles, hospital departments, diseases and symptoms for fine-grained analysis on language usage pattern and clinical diagnosis. We evaluate the performance of medical response generation, department routing and doctor recommendation on RealMedDial. Results show that RealMedDial are applicable to a wide range of NLP tasks with respect to medical dialogue.

Patients with low health literacy usually have difficulty understanding medical jargon and the complex structure of professional medical language. Although some studies are proposed to automatically translate expert language into layperson-understandable language, only a few of them focus on both accuracy and readability aspects simultaneously in the clinical domain. Thus, simplification of the clinical language is still a challenging task, but unfortunately, it is not yet fully addressed in previous work. To benchmark this task, we construct a new dataset named MedLane to support the development and evaluation of automated clinical language simplification approaches. Besides, we propose a new model called DECLARE that follows the human annotation procedure and achieves state-of-the-art performance compared with eight strong baselines. To fairly evaluate the performance, we also propose three specific evaluation metrics. Experimental results demonstrate the utility of the annotated MedLane dataset and the effectiveness of the proposed model DECLARE.

Health risk prediction is a challenge task that aims to predict whether patients would suffer from a certain disease/condition in the near future based on their historical EHR data. Although existing approaches can achieve better performance, none of them can deal with the noise existing in the EHR data explicitly. In this paper, we hypothesize that automatically removing noise from EHR data should help the models further improve the performance. Correspondingly, we propose a novel model named MedSkim, which is able to automatically rule out irrelevant visits and codes by effectively skimming through the EHR data. In particular, the proposed model has a code selection module that can directly make a skipping decision to each individual diagnosis codes and then remove the target-irrelevant ones. A backward probing RNN (BPRNN) is designed to reversely process the EHR data and provide a coarse grained representation learning for visits. Besides, a forward skipping RNN (FSRNN) is proposed to read the EHR in a preceding way and dynamically select important visits and codes based on the results of previous two modules. Finally, the risk prediction module uses the output hidden states from FSRNN for generating the final representation to make predictions. Additionally, we also design an extra regularization term based on the skip rate of the model and combine it with standard cross entropy loss to train the model in an end-to-end setting. Experimental results show that MedSkim achieves the best performance on three real-world datasets compared with the state-of-the-art baselines in terms of PR-AUC, F1 and Cohen’s Kappa. Moreover, the ablation study and case study confirm that the proposed MedSkim is reasonable and effective for removing noise from EHR data1.1The source code of the proposed MedSkim is available at https://github.com/SH-Src/MedSkim

The knowledge concept prerequisites describing the dependencies are critical for fundamental tasks such as material recommendations and there are a huge amount of concepts in Massive Open Online Courses (MOOCs). Thus it is necessary to develop automatic prerequisite relation annotation methods. Recently, a few methods have shown their effectiveness in discovering knowledge concept prerequisites in Moocs automatically. However, they suffer from two common issues, i.e., knowledge concepts are not thoroughly learnt, and informative supervision sources are ignored. To overcome these issues, we propose an end-to-end framework to incorporate the rich heterogeneous information in MOOCs, including the semantic, contextual and structural information of the learning materials as well as student video watching behaviors. Such useful information is not only used to derive entity representations but also as supervision to improve the prerequisite learning task. Experimental results on two public datasets show that the proposed framework outperforms state-of-the-art baselines in terms of precision, recall and F1 values and improves up to 9% in terms of F1 metrics. Besides, ablation study demonstrates the effectiveness of the proposed framework.

Generating text adversarial examples in the hard-label setting is a more realistic and challenging black-box adversarial attack problem, whose challenge comes from the fact that gradient cannot be directly calculated from discrete word replacements. Consequently, the effectiveness of gradient-based methods for this problem still awaits improvement. In this paper, we propose a gradient-based optimization method named LeapAttack to craft high-quality text adversarial examples in the hard-label setting. To specify, LeapAttack employs the word embedding space to characterize the semantic deviation between the two words of each perturbed substitution by their difference vector. Facilitated by this expression, LeapAttack gradually updates the perturbation direction and constructs adversarial examples in an iterative round trip: firstly, the gradient is estimated by transforming randomly sampled word candidates to continuous difference vectors after moving the current adversarial example near the decision boundary; secondly, the estimated gradient is mapped back to a new substitution word based on the cosine similarity metric. Extensive experimental results show that in the general case LeapAttack can efficiently generate high-quality text adversarial examples with the highest semantic similarity and the lowest perturbation rate in the hard-label setting.

Online learning platforms that can recommend tailored materials for different students have become increasingly popular recently. To enable personalized learning, it is critical and essential to automatically estimate the mastery levels of students, which motivates a new task in the education field, i.e., the student learning outcome prediction. Although several models have been proposed, most of them ignore the relations between questions and knowledge concepts. However, manually labeling the relations among questions only by experts is inefficient and impractical, due to the large volume of questions in the online learning platforms. Thus, an automatic inference of such relations is needed. In addition, different students may use different concepts when answering the same question, which makes the inferred relation between a question and concepts differ among students. To address these challenges, we propose to leverage information from a textbook to link questions with knowledge concepts that a student may retrieve. Correspondingly, we propose a novel framework named TESLOP, which can effectively utilize both textual and structural information in the textbook for student learning outcome prediction. The model simulates the process of a student picking an answer by recalling the knowledge obtained from the textbook and utilizing the knowledge to pick a correct answer. Experimental results show that the proposed TESLOP framework outperforms state-of-the-art baselines, which confirms the importance of leveraging textbook information in the student learning outcome prediction task. It also demonstrates that the proposed way of integrating such information is effective.

Human authentication is critical to protect personal and property security. Existing contactless authentication methods face some drawbacks, such as requiring large data size and low accuracy in cross-domain recognition, which hinders widespread popularization in practical applications. In this paper, we design and implement WiLCA, a WiFi-based lightweight contactless authentication system. First, we devise a Channel State Information (CSI) stream selection scheme to extract human movement features and reduce the sample size in the recognition process. Then, an AGO model is proposed, in which a Siamese Neural Network (SNN) framework with a cross-entropy module is used to guarantee accurate human authentication with limited data, and a lightweight GhostNet accelerates authentication with cheap operations. At last, extensive experiments are conducted to demonstrate the advantages of WiLCA, revealing that compared with state-of-the-art methods, WiLCA can reduce the data size by at least 2.5 x and achieve accurate authentication with an accuracy of over 98%.

Session-based recommendation aims to predict items that an anonymous user would like to purchase based on her short behavior sequence. The current approaches towards session-based recommendation only focus on modeling users' interest preferences, while they all ignore a key attribute of an item, i.e., the price. Many marketing studies have shown that the price factor significantly influences users' behaviors and the purchase decisions of users are determined by both price and interest preferences simultaneously. However, it is nontrivial to incorporate price preferences for session-based recommendation. Firstly, it is hard to handle heterogeneous information from various features of items to capture users' price preferences. Secondly, it is difficult to model the complex relations between price and interest preferences in determining user choices.

Audio video scene-aware dialog (AVSD) is a new but more challenging visual question answering (VQA) task because of the higher complexity of feature extraction and fusion brought by the additional modalities. Although recent methods have achieved early success in improving feature extraction technique for AVSD, the technique of feature fusion still needs further investigation. In this paper, inspired by the success of self-attention mechanism and the importance of understanding questions for VQA answering, we propose a question-guided self-attentive multi-modal fusion network (QUALIFIER) to improve the AVSD practice in the stage of feature fusion and answer generation. Specifically, after extracting features and learning a comprehensive feature for each modality, we first use the designed self-attentive multi-modal fusion (SMF) module to aggregate each feature with the correlated information learned from others. Later, by prioritizing the question feature, we concatenate it with each fused feature to guide the generation of a natural language response to the question. As for experimental results, QUALIFIER shows better performance than other baseline methods in the large-scale AVSD dataset named DSTC7. Additionally, the human evaluation and ablation study results also demonstrate the effectiveness of our network architecture.

Providing explanations for deep neural networks (DNNs) is essential for their use in domains wherein the interpretability of decisions is a critical prerequisite. Despite the plethora of work on interpreting DNNs, most existing solutions offer interpretability in an ad hoc, one-shot, and static manner, without accounting for the perception, understanding, or response of end-users, resulting in their poor usability in practice. In this paper, we argue that DNN interpretability should be implemented as the interactions between users and models. We present i-Algebra, a first-of-its-kind interactive framework for interpreting DNNs. At its core is a library of atomic, composable operators, which explain model behaviors at varying input granularity, during different inference stages, and from distinct interpretation perspectives. Leveraging a declarative query language, users are enabled to build various analysis tools (e.g., ``drill-down'', ``comparative'', ``what-if'' analysis) via flexibly composing such operators. We prototype i-Algebra and conduct user studies in a set of representative analysis tasks, including inspecting adversarial inputs, resolving model inconsistency, and cleansing contaminated data, all demonstrating its promising usability.

Xingyi Yang, Muchao Ye, Quanzeng You, Fenglong Ma. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Federated Learning has shown great potentials for the distributed data utilization and privacy protection. Most existing federated learning approaches focus on the supervised setting, which means all the data stored in each client has labels. However, in real-world applications, the client data are impossible to be fully labeled. Thus, how to exploit the unlabeled data should be a new challenge for federated learning. Although a few studies are attempting to overcome this challenge, they may suffer from information leakage or misleading information usage problems. To tackle these issues, in this paper, we propose a novel federated semi-supervised learning method named FedTriNet, which consists of two learning phases. In the first phase, we pre-train FedTriNet using labeled data with FedAvg. In the second phase, we aim to make most of the unlabeled data to help model learning. In particular, we propose to use three networks and a dynamic quality control mechanism to generate high-quality pseudo labels for unlabeled data, which are added to the training set. Finally, FedTriNet uses the new training set to retrain the model. Experimental results on three publicly available datasets show that the proposed FedTriNet outperforms state-of-the-art baselines under both IID and Non-IID settings.

The broad adoption of electronic health record (EHR) systems and the advances of deep learning technology have motivated the development of health risk prediction models, which mainly depend on the expressiveness and temporal modeling capacity of deep neural networks (DNNs) to improve prediction performance. Some further augment the prediction by using external knowledge, however, a great deal of EHR information inevitably loses during the knowledge mapping. In addition, prediction made by existing models usually lacks reliable interpretation, which undermines their reliability in guiding clinical decision-making. To solve these challenges, we propose MedRetriever, an effective and flexible framework that leverages unstructured medical text collected from authoritative websites to augment health risk prediction as well as to provide understandable interpretation. Besides, MedRetriever explicitly takes the target disease documents into consideration, which provide key guidance for the model to learn in a target-driven direction, i.e., from the target disease to the input EHR. To specify, MedRetriever can flexibly choose its backbone from major predictive models to learn the EHR embedding for each visit. After that, the EHR embedding and features of target disease documents are aggregated into a query by self-attention to retrieve highly relevant text segments from the medical text pool, which is stored in the dynamically updated text memory. Finally, the comprehensive EHR embedding and the text memory are used for prediction and interpretation. We evaluate MedRetriever against nine state-of-the-art approaches across three real-world EHR datasets, which consistently achieves the best performance in AUC and recall metrics and outperforms the best baseline by at least 4.8% in recall on three test datasets. Furthermore, we conduct case studies to show the easy-to-understand interpretation by MedRetriever.

Medication recommendation is an essential task of AI for healthcare. Existing works focused on recommending drug combinations for patients with complex health conditions solely based on their electronic health records. Thus, they have the following limitations: (1) some important data such as drug molecule structures have not been utilized in the recommendation process. (2) drug-drug interactions (DDI) are modeled implicitly, which can lead to sub-optimal results. To address these limitations, we propose a DDI-controllable drug recommendation model named SafeDrug to leverage drugs’ molecule structures and model DDIs explicitly. SafeDrug is equipped with a global message passing neural network (MPNN) module and a local bipartite learning module to fully encode the connectivity and functionality of drug molecules. SafeDrug also has a controllable loss function to control DDI level in the recommended drug combinations effectively. On a benchmark dataset, our SafeDrug is relatively shown to reduce DDI by 19.43% and improves 2.88% on Jaccard similarity between recommended and actually prescribed drug combinations over previous approaches. Moreover, SafeDrug also requires much fewer parameters than previous deep learning based approaches, leading to faster training by about 14% and around 2× speed-up in inference.

Thanks to the explosion of heterogeneous healthcare data and advanced machine learning and data mining techniques, specifically deep learning methods, we now have an opportunity to make difference in healthcare. In this tutorial, we will present state-of-the-art deep learning methods and their real-world applications, specifically focusing on exploring the unique characteristics of different types of healthcare data. The first half will be spent on introducing recent advances in mining structured healthcare data, including computational phenotyping, disease early detection/risk prediction and treatment recommendation. In the second half, we will focus on challenges specific to the unstructured healthcare data, and introduce advanced deep learning methods in automated ICD coding, understandable medical language translation, clinical trial mining, and medical report generation. This tutorial is intended for students, engineers and researchers who are interested in applying deep learning methods to healthcare, and prerequisite knowledge will be minimal. The tutorial will be concluded with open problems and a Q&A session.

Fake news travels at unprecedented speeds, reaches global audiences and puts users and communities at great risk via social media platforms. Deep learning based models show good performance when trained on large amounts of labeled data on events of interest, whereas the performance of models tends to degrade on other events due to domain shift. Therefore, significant challenges are posed for existing detection approaches to detect fake news on emergent events, where large-scale labeled datasets are difficult to obtain. Moreover, adding the knowledge from newly emergent events requires to build a new model from scratch or continue to fine-tune the model, which can be challenging, expensive, and unrealistic for real-world settings. In order to address those challenges, we propose an end-to-end fake news detection framework named MetaFEND, which is able to learn quickly to detect fake news on emergent events with a few verified posts. Specifically, the proposed model integrates meta-learning and neural process methods together to enjoy the benefits of these approaches. In particular, a label embedding module and a hard attention mechanism are proposed to enhance the effectiveness by handling categorical information and trimming irrelevant posts. Extensive experiments are conducted on multimedia datasets collected from Twitter and Weibo. The experimental results show our proposed MetaFEND model can detect fake news on never-seen events effectively and outperform the state-of-the-art methods.

Predicting the current learning outcome of a student based on his/her responses to previous questions is a vital task for personalized education.Deep learning based models have achieved satisfactory performance on this task as they can automatically extract meaningful signals related to student learning outcome predictions.However, these models are unable to leverage the following critical pieces of information.First, the correct choice to a question is often accompanied by a textual explanation on why this choice is correct.This explanation text contains valuable informa-

Successful health risk prediction demands accuracy and reliability of the model. Existing predictive models mainly depend on mining electronic health records (EHR) with advanced deep learning techniques to improve model accuracy. However, they all ignore the importance of publicly available online health data, especially socioeconomic status, environmental factors, and detailed demographic information for each location, which are all strong predictive signals and can definitely augment precision medicine. To achieve model reliability, the model needs to provide accurate prediction and uncertainty score of the prediction. However, existing uncertainty estimation approaches often failed in handling high-dimensional data, which are present in multi-sourced data.

The broad adoption of electronic health records (EHR) data and the availability of biomedical knowledge graphs (KGs) on the web have provided clinicians and researchers unprecedented resources and opportunities for conducting health risk predictions to improve healthcare quality and medical resource allocation. Existing methods have focused on improving the EHR feature representations using attention mechanisms, time-aware models, or external knowledge. However, they ignore the importance of using personalized information to make predictions. Besides, the reliability of their prediction interpretations needs to be improved since their interpretable attention scores are not explicitly reasoned from disease progression paths. In this paper, we propose MedPath to solve these challenges and augment existing risk prediction models with the ability to use personalized information and provide reliable interpretations inferring from disease progression paths. Firstly, MedPath extracts personalized knowledge graphs (PKGs) containing all possible disease progression paths from observed symptoms to target diseases from a large-scale online medical knowledge graph. Next, to augment existing EHR encoders for achieving better predictions, MedPath learns a PKG embedding by conducting multi-hop message passing from symptom nodes to target disease nodes through a graph neural network encoder. Since MedPath reasons disease progression by paths existing in PKGs, it can provide explicit explanations for the prediction by pointing out how observed symptoms can finally lead to target diseases. Experimental results on three real-world medical datasets show that MedPath is effective in improving the performance of eight state-of-the-art methods with higher F1 scores and AUCs. Our case study also demonstrates that MedPath can greatly improve the explicitness of the risk prediction interpretation.1

Today social media has become the primary source for news. Via social media platforms, fake news travel at unprecedented speeds, reach global audiences and put users and communities at great risk. Therefore, it is extremely important to detect fake news as early as possible. Recently, deep learning based approaches have shown improved performance in fake news detection. However, the training of such models requires a large amount of labeled data, but manual annotation is time-consuming and expensive. Moreover, due to the dynamic nature of news, annotated samples may become outdated quickly and cannot represent the news articles on newly emerged events. Therefore, how to obtain fresh and high-quality labeled samples is the major challenge in employing deep learning models for fake news detection. In order to tackle this challenge, we propose a reinforced weakly-supervised fake news detection framework, i.e., WeFEND, which can leverage users' reports as weak supervision to enlarge the amount of training data for fake news detection. The proposed framework consists of three main components: the annotator, the reinforced selector and the fake news detector. The annotator can automatically assign weak labels for unlabeled news based on users' reports. The reinforced selector using reinforcement learning techniques chooses high-quality samples from the weakly labeled data and filters out those low-quality ones that may degrade the detector's prediction performance. The fake news detector aims to identify fake news based on the news content. We tested the proposed framework on a large collection of news articles published via WeChat official accounts and associated user reports. Extensive experiments on this dataset show that the proposed WeFEND model achieves the best performance compared with the state-of-the-art methods.

Recent advances in information extraction have motivated the automatic construction of huge Knowledge Graphs (KGs) by mining from large-scale text corpus. However, noisy facts are unavoidably introduced into KGs that could be caused by automatic extraction.

Risk prediction using electronic health records (EHR) is a challenging data mining task due to the two-level hierarchical structure of EHR data. EHR data consist of a set of time-ordered visits, and within each visit, there is a set of unordered diagnosis codes. Existing approaches focus on modeling temporal visits with deep neural network (DNN) techniques. However, they ignore the importance of modeling diagnosis codes within visits, and a lot of task-unrelated information within visits usually leads to unsatisfactory performance of existing approaches. To minimize the effect caused by noise information of EHR data, in this paper, we propose a novel DNN for risk prediction termed as LSAN, which consists of a Hierarchical Attention Module (HAM) and a Temporal Aggregation Module (TAM). Particularly, LSAN applies HAM to model the hierarchical structure of EHR data. Using the attention mechanism in the hierarchy of diagnosis code, HAM is able to retain diagnosis details and assign flexible attention weights to different diagnosis codes by their relevance to corresponding diseases. Moreover, the attention mechanism in the hierarchy of visit learns a comprehensive feature throughout the visit history by paying greater attention to visits with higher relevance. Based on the foundation laying by HAM, TAM uses a two-pathway structure to learn a robust temporal aggregation mechanism among all visits for LSAN. It extracts long-term dependencies by a Transformer encoder and short-term correlations by a parallel convolutional layer among different visits. With the construction of HAM and TAM, LSAN achieves the state-of-the-art performance on three real-world datasets with larger AUCs, recalls and F1 scores. Furthermore, the model analysis results demonstrate the effectiveness of the network construction with good interpretability and robustness of decision making by LSAN.

Handwriting recognition system provides people a convenient and alternative way for writing in the air with fingers rather than typing keyboards. For people with blurred vision and patients with generalized hand neurological disease, writing in the air is particularly attracting due to the small input screen of smartphones and smartwatches. Existing recognition systems still face drawbacks such as requiring to wear dedicated devices, relatively low accuracy and infeasible for cross domain identification, which greatly limit the usability of these systems. To address these issues, we propose WiWrite, an accurate device-free handwriting recognition system which allows writing in the air without a need of attaching any device to the user. Specifically, we use Commercial Off-The-Shelf (COTS) WiFi hardware to achieve fine-grained finger tracking. We develop a CSI division scheme to process the noisy raw WiFi channel state information (CSI), which stabilizes the CSI phase and reduces the noise of the CSI amplitude. To automatically retain low noise data for identification, we propose a self-paced dense convolutional network (SPDCN), which consists of the self-paced loss function based on a modified convolutional neural network, together with a dense convolutional network. Comprehensive experiments are conducted to show the merits of WiWrite, revealing that, the recognition accuracies for the same-size input and different-size input are 93.6% and 89.0%, respectively. Moreover, WiWrite can achieve a one-fit-for-all recognition regardless of environment diversities.

Outlier detection is of vital importance for various fields and applications. Existing works mainly focus on identifying outliers from underlying datasets, while how to provide sense-making explanations is largely ignored. In this paper, we propose to visualize data points in a set of scatter plots on two-dimensional (2-D) feature spaces that can provide meaningful explanations about the outlying behavior of outliers. Data are typically multidimensional and the number of 2-D combinations could be huge. Also, outliers may have diverse characteristics, and thus the global scatter plots containing all of outliers may degrade the explanation effectiveness for those outliers having idiosyncratic abnormal 2-D spaces. To address this problem, we propose a new outlier explanation approach, called LP-Explain, which tries to identify the set of best Local Pictorial explanations (defined as the scatter plots in the 2-D space of the feature pairs) that can Explain the behavior for cluster of outliers. We first define an effective measure to quantify the similarity between outliers, and then cluster outliers into different groups based on their abnormal feature pairs. We then propose to weigh the importance of feature pairs within each cluster through a multi-task learning framework to select the set of top feature pairs that best explain various outlier clusters. By adjusting a user-defined parameter indicating the “localization level”, the proposed method can attain both global and local results for the explanation of the outliers. 2-D visual explanations can be plotted for the top-weighted feature pairs of each cluster. We conduct experiments on various public datasets, which show that the proposed approach can provide more meaningful explanations about the outlying behavior in a dataset.

To provide accurate and explainable misinformation detection, it is often useful to take an auxiliary source (e.g., social context and knowledge base) into consideration. Existing methods use social contexts such as users' engagements as complementary information to improve detection performance and derive explanations. However, due to the lack of sufficient professional knowledge, users seldom respond to healthcare information, which makes these methods less applicable. In this work, to address these shortcomings, we propose a novel knowledge guided graph attention network for detecting health misinformation better. Our proposal, named as DETERRENT, leverages on the additional information from medical knowledge graph by propagating information along with the network, incorporates a Medical Knowledge Graph and an Article-Entity Bipartite Graph, and propagates the node embeddings through Knowledge Paths. In addition, an attention mechanism is applied to calculate the importance of entities to each article, and the knowledge guided article embeddings are used for misinformation detection. DETERRENT addresses the limitation on social contexts in the healthcare domain and is capable of providing useful explanations for the results of detection. Empirical validation using two real-world datasets demonstrated the effectiveness of DETERRENT. Comparing with the best results of eight competing methods, in terms of F1 Score, DETERRENT outperforms all methods by at least 4.78% on the diabetes dataset and 12.79% on cancer dataset. We release the source code of DETERRENT at: https://github.com/cuilimeng/DETERRENT.

Deep learning methods especially recurrent neural network based models have demonstrated early success in disease risk prediction on longitudinal patient data. Existing works follow a strong assumption to implicitly assume the stationary disease progression during each time period, and thus, take a homogeneous way to decay the information from previous time steps for all patients. However,in reality, disease progression is non-stationary. Besides, the key time steps for a target disease vary among patients. To leverage time information for risk prediction in a more reasonable way, we propose a new hierarchical time-aware attention network, named HiTANet, which imitates the decision making process of doctors inrisk prediction. Particularly, HiTANet models time information in local and global stages. The local evaluation stage has a time aware Transformer that embeds time information into visit-level embed-ding and generates local attention weight for each visit. The global synthesis stage further adopts a time-aware key-query attention mechanism to assign global weights to different time steps. Finally, the two types of attention weights are dynamically combined to generate the patient representations for further risk prediction. We evaluate HiTANet on three real-world datasets. Compared with the best results among twelve competing baselines, HiTANet achieves over 7% in terms of F1 score on all datasets, which demonstrates the effectiveness of the proposed model and the necessity of modeling time information in risk prediction task.

Evasion attack on discrete data is a challenging, while practically interesting research topic. It is intrinsically an NP-hard combinatorial optimization problem. Characterizing the conditions guaranteeing the solvability of an evasion attack task thus becomes the key to understand the adversarial threat. Our study is inspired by the weak submodularity theory. We characterize the attackability of a targeted classifier on discrete data in evasion attack by bridging the attackability measurement and the regularity of the targeted classifier. Based on our attackability analysis, we propose a computationally efficient orthogonal matching pursuit-guided attack method for evasion attack on discrete data. It provides provably computational efficiency and attack performances. Substantial experimental results on real-world datasets validate the proposed attackability conditions and the effectiveness of the proposed attack method.

Predicting diseases for patients is an important and practical task in healthcare informatics. Existing disease prediction models focus on common diseases, i.e., there are enough available EHR data and prior medical knowledge for analyzing them. However, those models may not work for rare disease prediction as it is extremely hard to collect enough EHR data with such diseases. To tackle these issues, in this paper, we design a novel rare disease prediction system, which not only generates EHR data but also automatically selects high-quality generated data to further improve the predictive performance. Three components are designed in the system: data generation, data selection, and prediction. In particular, we propose MaskEHR to generate diverse EHR data based on the data from patients suffering from the given diseases. To remove noise information in the generated EHR data, we further design a reinforcement learning-based data selector, called RL-Selector, which can automatically choose the high-quality generated EHR data. Finally, the prediction component is used to identify patients who will potentially suffer the given diseases. These three components work together and enhance each other. Experiments on three real healthcare datasets show that the proposed system outperforms existing approaches on rare disease prediction task.

Digital screens, such as liquid crystal displays (LCDs), are vulnerable to attacks (e.g., "shoulder surfing") that can bypass security protection services (e.g., firewall) to steal confidential information from intended victims. The conventional practice to mitigate these threats is isolation. An isolated zone, without accessibility, proximity, and line-of-sight, seems to bring personal devices to a truly secure place.In this paper, we revisit this historical topic and re-examine the security risk of screen attacks in an isolation scenario mentioned above. Specifically, we identify and validate a new and practical side-channel attack for screen content via liquid crystal nematic state estimation using a low-cost radio-frequency sensor. By leveraging the relationship between the screen content and the states of liquid crystal arrays in displays, we develop WaveSpy, an end-to-end portable through-wall screen attack system. WaveSpy comprises a low-cost, energy-efficient and light-weight millimeter-wave (mmWave) probe which can remotely collect the liquid crystal state response to a set of mmWave stimuli and facilitate screen content inference, even when the victim’s screen is placed in an isolated zone. We intensively evaluate the performance and practicality of WaveSpy in screen attacks, including over 100 different types of content on 30 digital screens of modern electronic devices. WaveSpy achieves an accuracy of 99% in screen content type recognition and a success rate of 87.77% in Top-3 sensitive information retrieval under real-world scenarios, respectively. Furthermore, we discuss several potential defense mechanisms to mitigate screen eavesdropping similar to WaveSpy.

In the era of big data, it is easy for us collect a huge number of image and text data. However, we frequently face the real-world problems with only small (labeled) data in some domains, such as healthcare and urban computing. The challenge is how to make machine learn algorithms still work well with small data? To solve this challenge, in this tutorial, we will cover the state-of-the-art machine learning techniques to handle small data issue. In particular, we focus on the following three aspects: (1) Providing a comprehensive review of recent advances in exploring the power of knowledge transfer, especially focusing on meta-learning; (2) introducing the cutting-edge techniques of incorporating human/expert knowledge into machine learning models; and (3) identifying the open challenges to data augmentation techniques, such as generative adversarial networks. We believe this is an emerging and potentially high-impact topic in computational data science, which will attract both researchers and practitioners from academia and industry.

This paper presents a novel framework, MGNER, for Multi-Grained Named Entity Recognition where multiple entities or entity mentions in a sentence could be nonoverlapping or totally nested.Different from traditional approaches regarding NER as a sequential labeling task and annotate entities consecutively, MGNER detects and recognizes entities on multiple granularities: it is able to recognize named entities without explicitly assuming non-overlapping or totally nested structures.MGNER consists of a Detector that examines all possible word segments and a Classifier that categorizes entities.In addition, contextual information and a self-attention mechanism are utilized throughout the framework to improve the NER performance.Experimental results show that MGNER outperforms current state-of-the-art baselines up to 4.4% in terms of the F1 score among nested/non-overlapping NER tasks.* Work was done when the author Yaliang Li was at Tencent America.

With the popular use of mobile devices, it becomes increasingly important to conduct analysis on distributed data collected from multiple devices. Federated learning is a distributed learning framework which takes advantage of the training data and computational ability of scattered mobile devices to learn prediction models, and multi-task learning infers personalized but shared models among devices. Some recent work has integrated federated and multi-task learning, but such approaches may be impractical and inefficient in the online scenario, e.g., when new mobile devices keep joining the mobile computing system. To address this challenge, we propose OFMTL, an online federated multi-task learning algorithm, which learns the model parameters for the new device without revisiting the data of existing devices. The model parameters are derived by an effective way that combines the information inferred from local data and information borrowed from existing models. Through extensive experiments on three real datasets, we show that the proposed OFMTL framework achieves comparable accuracy to the existing algorithms but with much smaller computation, transmission and storage cost.

Crowdsourcing has become a popular way to obtain a large volume of labeled data due to its low cost and high efficiency. Usually the crowdsourcing process enables redundancy in the collected labels in order to ensure the correctness of item labels. However, workers on the crowdsourcing platform may make mistakes on some items, leading to inconsistent labels. In this case, it is important to aggregate these noisy labels and obtain the true labels of the items. The correctness of the item label provided by a worker depends on both the worker's ability and the property of the item. However, most of the existing models consider the effect of workers' abilities but ignore that of the item properties. In this paper, we propose a novel crowdsourcing aggregation method (IProWA) which incorporates the modeling of not only worker expertise level but also item property. In particular, items are represented by a K dimensional vector (i.e., item parameter), where K is the number of possible categories and each dimension represents a category. The proposed model transforms the true label estimation into the estimation of item parameters as it connects the true label and the parameters of an item. In worker modeling, it models the different category propensities among different worker groups. Experimental results show that the performance of the proposed model is comparable to that of the state-of-the-art baselines and the learned item parameters can help interpret the property of that item.

Utilizing multiple modalities to learn a good distance metric is of vital importance for various clinical applications. However, it is common that modalities are incomplete for some patients due to various technical and practical reasons in healthcare datasets. Existing metric learning methods cannot directly learn the distance metric on such data with missing modalities. Nevertheless, the incomplete data contains valuable information to characterize patient similarity and modality relationships, and they should not be ignored during the learning process. To tackle the aforementioned challenges, we propose a metric learning framework to perform missing modality completion and multi-modal metric learning simultaneously. Employing the generative adversarial networks, we incorporate both complete and incomplete data to learn the mapping relationship between modalities. After completing the missing modalities, we use the nonlinear representations extracted by the discriminator to learn the distance metric among patients. Through jointly training the adversarial generation part and metric learning, the similarity among patients can be learned on data with missing modalities. Experimental results show that the proposed framework learns more accurate distance metric on real-world healthcare datasets with incomplete modalities, comparing with the state-of-the-art approaches. Meanwhile, the quality of the generated modalities can be preserved.

The discovery of Wireless Power Transfer (WPT) technologies makes charging more convenient and reliable. Among all the existing WPT technologies, directional WPT is more efficient and has been successfully applied to supply energy for wireless rechargeable sensor networks (WRSNs). However, the state-of-the-art methods ignore the anisotropic energy receiving property of rechargeable sensors, resulting in energy wastage. In order to address this issue, in this paper, we point out that the received energy of a sensor is not only relative to the distance, but also relative to the angle between the sensor and the charger's orientation in directional WPT. Towards this end, we derive a pragmatic energy transfer model verified by experiments. In particular, we focus on a Minimal chArging Delay (MAD) problem to reduce charging delays. To obtain the optimal solution, we formulate the problem as a linear programming problem. Moreover, we introduce a method of charging power discretization, which significantly reduces the search space and bounds the performance gap to the optimal one with a 1/1-ϵ2approximation ratio. Besides, a merging method is introduced for a more practical application scenario. Finally, we demonstrate that our methods outperform the Set Cover baseline method by an average of 34.2% through simulations and experiments.

The increasing need for labeled data has brought the booming growth of crowdsourcing in a wide range of high-impact real-world applications, such as collaborative knowledge (e.g., data annotations, language translations), collective creativity (e.g., analogy mining, crowdfunding), and reverse Turing test (e.g., CAPTCHA-like systems), etc. In the context of supervised learning, crowdsourcing refers to the annotation procedure where the data items are outsourced and processed by a group of mostly unskilled online workers. Thus, the researchers or the organizations are able to collect large amount of information via the feedback of the crowd in a short time with a low cost.

In recent years, significant research efforts have been spent towards building intelligent and user-friendly IoT systems to enable a new generation of applications capable of performing complex sensing and recognition tasks. In many of such applications, there are usually multiple different sensors monitoring the same object. Each of these sensors can be regarded as an information source and provides us a unique "view" of the observed object. Intuitively, if we can combine the complementary information carried by multiple sensors, we will be able to improve the sensing performance. Towards this end, we propose DeepFusion, a unified multi-sensor deep learning framework, to learn informative representations of heterogeneous sensory data. DeepFusion can combine different sensors' information weighted by the quality of their data and incorporate cross-sensor correlations, and thus can benefit a wide spectrum of IoT applications. To evaluate the proposed DeepFusion model, we set up two real-world human activity recognition testbeds using commercialized wearable and wireless sensing devices. Experiment results show that DeepFusion can outperform the state-of-the-art human activity recognition methods.

Relation classification is a basic yet important task in natural language processing. Existing relation classification approaches mainly rely on distant supervision, which assumes that a bag of sentences mentioning a pair of entities and extracted from a given corpus should express the same relation type of this entity pair. The training of these models needs a lot of high-quality bag-level data. However, in some specific domains, such as medical domain, it is difficult to obtain sufficient and high-quality sentences in a text corpus that mention two entities with a certain medical relation between them. In such a case, it is hard for existing discriminative models to capture the representative features (i.e., common patterns) from diversely expressed entity pairs with a given relation. Thus, the classification performance cannot be guaranteed when limited features are obtained from the corpus. To address this challenge, in this paper, we propose to employ a generative model, called conditional variational autoencoder (CVAE), to handle the pattern sparsity. We define that each relation has an individually learned latent distribution from all possible sentences expressing this relation. As these distributions are learned based on the purpose of input reconstruction, the model's classification ability may not be strong enough and should be improved. By distinguishing the differences among different relation distributions, a margin-based regularizer is designed, which leads to a margin-based CVAE (MCVAE) that can significantly enhance the classification ability. Besides, MCVAE can automatically generate semantically meaningful patterns that describe the given relations. Experiments on two real-world datasets validate the effectiveness of the proposed MCVAE on the tasks of relation classification and relation-specific pattern generation.

BACKGROUND: Sleep is a complex and dynamic biological process characterized by different sleep patterns. Comprehensive sleep monitoring and analysis using multivariate polysomnography (PSG) records has achieved significant efforts to prevent sleep-related disorders. To alleviate the time consumption caused by manual visual inspection of PSG, automatic multivariate sleep stage classification has become an important research topic in medical and bioinformatics. RESULTS: We present a unified hybrid self-attention deep learning framework, namely HybridAtt, to automatically classify sleep stages by capturing channel and temporal correlations from multivariate PSG records. We construct a new multi-view convolutional representation module to learn channel-specific and global view features from the heterogeneous PSG inputs. The hybrid attention mechanism is designed to further fuse the multi-view features by inferring their dependencies without any additional supervision. The learned attentional representation is subsequently fed through a softmax layer to train an end-to-end deep learning model. CONCLUSIONS: We empirically evaluate our proposed HybridAtt model on a benchmark PSG dataset in two feature domains, referred to as the time and frequency domains. Experimental results show that HybridAtt consistently outperforms ten baseline methods in both feature spaces, demonstrating the effectiveness of HybridAtt in the task of sleep stage classification.

The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. The proposed framework is further adapted to deal with large-scale truth discovery task in distributed paradigm. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.

Detecting local events (e.g., protests, accidents) in real-time is an important task needed by a wide spectrum of real-world applications. In recent years, with the proliferation of social media platforms, we can access massive geo- tagged social messages, which can serve as a precious resource for timely local event detection. However, existing local event detection methods either suffer from unsatisfactory performances or need intensive annotations. These limitations make existing methods impractical for large-scale applications. Through the analysis of real-world datasets, we found that the informativeness level of social media users, which is neglected by existing work, plays a highly critical role in distilling event-related information from noisy social media contexts. Motivated by this finding, we propose an unsupervised framework, named LEDetect, to estimate the informativeness level of social media users and leverage the power of highly informative users for local event detection. Experiments on a large-scale real-world dataset show that the proposed LEDetect model can improve the performance of event detection compared with the state-of-the-art unsupervised approach. Also, we use case studies to show that the events discovered by the proposed model are of high quality and the extracted highly informative users are reasonable.

The goal of diagnosis prediction task is to predict the future health information of patients from their historical Electronic Healthcare Records (EHR). The most important and challenging problem of diagnosis prediction is to design an accurate, robust and interpretable predictive model. Existing work solves this problem by employing recurrent neural networks (RNNs) with attention mechanisms, but these approaches suffer from the data sufficiency problem. To obtain good performance with insufficient data, graph-based attention models are proposed. However, when the training data are sufficient, they do not offer any improvement in performance compared with ordinary attention-based models. To address these issues, we propose KAME, an end-to-end, accurate and robust model for predicting patients' future health information. KAME not only learns reasonable embeddings for nodes in the knowledge graph, but also exploits general knowledge to improve the prediction accuracy with the proposed knowledge attention mechanism. With the learned attention weights, KAME allows us to interpret the importance of each piece of knowledge in the graph. Experimental results on three real world datasets show that the proposed KAME significantly improves the prediction performance compared with the state-of-the-art approaches, guarantees the robustness with both sufficient and insufficient data, and learns interpretable disease representations.

A clinically meaningful distance metric, which is learned from measuring patient similarity, plays an important role in clinical decision support applications. Several metric learning approaches have been proposed to measure patient similarity, but they are mostly designed for learning the metric at only one time point/interval. It leads to a problem that those approaches cannot reflect the similarity variations among patients with the progression of diseases. In order to capture similarity information from multiple future time points simultaneously, we formulate a multi-task metric learning approach to identify patient similarity. However, it is challenging to directly apply traditional multi-task metric learning methods to learn such similarities due to the high dimensional, complex and noisy nature of healthcare data. Besides, the disease labels often have clinical relationships, which should not be treated as independent. Unfortunately, traditional formulation of the loss function ignores the degree of labels' similarity. To tackle the aforementioned challenges, we propose mtTSML, a multi-task triplet constrained sparse metric learning method, to monitor the similarity progression of patient pairs. In the proposed model, the distance for each task can be regarded as the combination of a common part and a task-specific one in the transformed low-rank space. We then perform sparse feature selection for each individual task to select the most discriminative information. Moreover, we use triplet constraints to guarantee the margin between similar and less similar pairs according to the ordered information of disease severity levels (i.e. labels). The experimental results on two real-world healthcare datasets show that the proposed multi-task metric learning method significantly outperforms the state-of-the-art baselines, including both single-task and multi-task metric learning methods.

A tensor (i.e., an N-mode array) is a natural representation for multidimensional data. Tucker Decomposition (TD) is one of the most popular methods, and a series of batch TD algorithms have been extensively studied and widely applied in signal/image processing, bioinformatics, etc. However, in many applications, the large-scale tensor is dynamically evolving at all modes, which poses significant challenges for existing approaches to track the TD for such dynamic tensors. In this paper, we propose an efficient Online Tucker Decomposition (eOTD) approach to track the TD of dynamic tensors with an arbitrary number of modes. We first propose corollaries on the multiplication of block tensor matrix. Based on this corollary, eOTD allows us 1) to update the projection matrices using those projection matrices from the previous timestamp and the auxiliary matrices from the current timestamp, and 2) to update the core tensor by a sum of tensors that are obtained by multiplying smaller tensors with matrices. The auxiliary matrices are obtained by solving a series of least square regression tasks, not by performing Singular Value Decompositions (SVD). This overcomes the bottleneck in computation and storage caused by computing SVDs on largescale data. A Modified Gram-Schmidt (MGS) process is further applied to orthonormalize the projection matrices. Theoretically, the output of the eOTD framework is guaranteed to be lowrank. We further prove that the MGS process will not increase Tucker decomposition error. Empirically, we demonstrate that the proposed eOTD achieves comparable accuracy with a significant speedup on both synthetic and real data, where the speedup can be more than 1,500 times on large-scale data.

Recent advances in attention networks have gained enormous interest in time series data mining. Various attention mechanisms are proposed to soft-select relevant timestamps from temporal data by assigning learnable attention scores. However, many real-world tasks involve complex multivariate time series that continuously measure target from multiple views. Different views may provide information of different levels of quality varied over time, and thus should be assigned with different attention scores as well. Unfortunately, the existing attention-based architectures cannot be directly used to jointly learn the attention scores in both time and view domains, due to the data structure complexity. Towards this end, we propose a novel multi-view attention network, namely MuVAN, to learn fine-grained attentional representations from multivariate temporal data. MuVAN is a unified deep learning model that can jointly calculate the two-dimensional attention scores to estimate the quality of information contributed by each view within different timestamps. By constructing a hybrid focus procedure, we are able to bring more diversity to attention, in order to fully utilize the multi-view information. To evaluate the performance of our model, we carry out experiments on three real-world benchmark datasets. Experimental results show that the proposed MuVAN model outperforms the state-of-the-art deep representation approaches in different real-world tasks. Analytical results through a case study demonstrate that MuVAN can discover discriminative and meaningful attention scores across views over time, which improves the feature representation of multivariate temporal data.

Predicting the risk of potential diseases from Electronic Health Records (EHR) has attracted considerable attention in recent years, especially with the development of deep learning techniques. Compared with traditional machine learning models, deep learning based approaches achieve superior performance on risk prediction task. However, none of existing work explicitly takes prior medical knowledge (such as the relationships between diseases and corresponding risk factors) into account. In medical domain, knowledge is usually represented by discrete and arbitrary rules. Thus, how to integrate such medical rules into existing risk prediction models to improve the performance is a challenge. To tackle this challenge, we propose a novel and general framework called PRIME for risk prediction task, which can successfully incorporate discrete prior medical knowledge into all of the state-of-the-art predictive models using posterior regularization technique. Different from traditional posterior regularization, we do not need to manually set a bound for each piece of prior medical knowledge when modeling desired distribution of the target disease on patients. Moreover, the proposed PRIME can automatically learn the importance of different prior knowledge with a log-linear model.Experimental results on three real medical datasets demonstrate the effectiveness of the proposed framework for the task of risk prediction

As news reading on social media becomes more and more popular, fake news becomes a major issue concerning the public and government. The fake news can take advantage of multimedia content to mislead readers and get dissemination, which can cause negative effects or even manipulate the public events. One of the unique challenges for fake news detection on social media is how to identify fake news on newly emerged events. Unfortunately, most of the existing approaches can hardly handle this challenge, since they tend to learn event-specific features that can not be transferred to unseen events. In order to address this issue, we propose an end-to-end framework named Event Adversarial Neural Network (EANN), which can derive event-invariant features and thus benefit the detection of fake news on newly arrived events. It consists of three main components: the multi-modal feature extractor, the fake news detector, and the event discriminator. The multi-modal feature extractor is responsible for extracting the textual and visual features from posts. It cooperates with the fake news detector to learn the discriminable representation for the detection of fake news. The role of event discriminator is to remove the event-specific features and keep shared features among events. Extensive experiments are conducted on multimedia datasets collected from Weibo and Twitter. The experimental results show our proposed EANN model can outperform the state-of-the-art methods, and learn transferable feature representations.

Truth discovery has attracted increasingly more attention due to its ability to distill trustworthy information from noisy multi-sourced data without any supervision. However, most existing truth discovery methods are designed for structured data, and cannot meet the strong need to extract trustworthy information from raw text data as text data has its unique characteristics. The major challenges of inferring true information on text data stem from the multifactorial property of text answers (i.e., an answer may contain multiple key factors) and the diversity of word usages (i.e., different words may have the same semantic meaning). To tackle these challenges, in this paper, we propose a novel truth discovery method, named "TextTruth", which jointly groups the keywords extracted from the answers of a specific question into multiple interpretable factors, and infers the trustworthiness of both answer factors and answer providers. After that, the answers to each question can be ranked based on the estimated trustworthiness of factors. The proposed method works in an unsupervised manner, and thus can be applied to various application scenarios that involve text data. Experiments on three real-world datasets show that the proposed TextTruth model can accurately select trustworthy answers, even when these answers are formed by multiple factors.

Driven by a wide range of real-world applications, significant efforts have recently been made to explore device-free human activity recognition techniques that utilize the information collected by various wireless infrastructures to infer human activities without the need for the monitored subject to carry a dedicated device. Existing device free human activity recognition approaches and systems, though yielding reasonably good performance in certain cases, are faced with a major challenge. The wireless signals arriving at the receiving devices usually carry substantial information that is specific to the environment where the activities are recorded and the human subject who conducts the activities. Due to this reason, an activity recognition model that is trained on a specific subject in a specific environment typically does not work well when being applied to predict another subject's activities that are recorded in a different environment. To address this challenge, in this paper, we propose EI, a deep-learning based device free activity recognition framework that can remove the environment and subject specific information contained in the activity data and extract environment/subject-independent features shared by the data collected on different subjects under different environments. We conduct extensive experiments on four different device free activity recognition testbeds: WiFi, ultrasound, 60 GHz mmWave, and visible light. The experimental results demonstrate the superior effectiveness and generalizability of the proposed EI framework.

Truth discovery, with the goal of inferring true information from massive data through aggregating the information from multiple data sources, has attracted significant attention in recent years. It has demonstrated great advantages in real applications since it can automatically learn the reliability degrees of the data sources without supervision and in turn helps to find more reliable information. In many applications, however, the data may arrive in a stream and present various temporal patterns. Unfortunately, there is no existing truth discovery work that can handle such time series data. To tackle this challenge, we propose a novel online truth discovery framework that incorporates the predictions on the time series data into the truth estimation process. By jointly considering the multi-source information and the temporal patterns of the time series data, the proposed framework can improve the accuracy of the truth discovery results as well as the time series prediction. The effectiveness of the proposed framework is validated on both synthetic and real-world datasets.

The ubiquitous and fine-grained features of WiFi signals make it promising for achieving device-free authentication. However, traditional methods suffer from drawbacks such as sensitivity to environmental dynamics, low accuracy, long delay, etc. In this paper, we introduce how to validate human identity using the ubiquitous WiFi signals. We develop WiAU, a device-free authentication system which only utilizes a Commodity Off-The-Shelf (COTS) router and a laptop. We describe the constitutions of WiAU and how it works in detail. Through collecting channel state information (CSI) profiles, WiAU automatically segments coherent activities and walking gait using an automatic segment algorithm (ASA). Then, a ResNet algorithm with two dedicated loss functions is designed to validate legal users and recognize illegal ones. Finally, experiments are conducted from different scenes to highlight the superiorities of WiAU in terms of high accuracy, short delay and robustness, revealing that WiAU has an accuracy of over 98% in recognizing human identity and human activities respectively.

Predicting patients' risk of developing certain diseases is an important research topic in healthcare. Personalized predictive modeling, which focuses on building specific models for individual patients, has shown its advantages on utilizing heterogeneous health data compared to global models trained on the entire population. Personalized predictive models use information from similar patient cohorts, in order to capture the specific characteristics. Accurately identifying and ranking the similarity among patients based on their historical records is a key step in personalized modeling. The electric health records (EHRs), which are irregular sampled and have varied patient visit lengths, cannot be directly used to measure patient similarity due to lack of an appropriate vector representation. In this paper, we build a novel time fusion CNN framework to simultaneously learn patient representations and measure pairwise similarity. Compared to a traditional CNN, our time fusion CNN can learn not only the local temporal relationships but also the contributions from each time interval. Along with the similarity learning process, the output information which is the probability distribution is used to rank similar patients. Utilizing the similarity scores, we perform personalized disease predictions, and compare the effect of different vector representations and similarity learning metrics.

In the big data era, the information about the same object collected from multiple sources is inevitably conflicting. The task of identifying true information (i.e., the truths) among conflicting data is referred to as truth discovery, which incorporates the estimation of source reliability degrees into the aggregation of multi-source data. However, in many real-world applications, large-scale data are distributed across multiple servers. Traditional truth discovery approaches cannot handle this scenario due to the constraints of communication overhead and privacy concern. Another limitation of most existing work is that they ignore the differences among objects, i.e., they treat all the objects equally. This limitation would be exacerbated in distributed environments where significant differences exist among the objects. To tackle the aforementioned issues, in this paper, we propose a novel distributed truth discovery framework (DTD), which can effectively and efficiently aggregate conflicting data stored across distributed servers, with the differences among the objects as well as the importance level of each server being considered. The proposed framework consists of two steps: the local truth computation step conducted by each local server and the central truth estimation step taking place in the central server. Specifically, we introduce the uncertainty values to model the differences among objects, and propose a new uncertainty-based truth discovery method (UbTD) for calculating the true information of objects in each local server. The outputs of the local truth computation step include the estimated local truths and the variances of objects, which are the input information of the central truth estimation step. To infer the final true information in the central server, we propose a new algorithm to aggregate the outputs of all the local servers with the quality of different local servers taken into account. The proposed distributed truth discovery framework can infer object truths without delivering any raw data to the central server, and thus can reduce communication overhead as well as preserve data privacy. Experimental results on three real world datasets show that the proposed DTD framework can efficiently estimate object truths with accuracy guarantee, and the proposed UbTD algorithm significantly outperforms the state-of-the-art batch truth discovery approaches.

Predicting the future health information of patients from the historical Electronic Health Records (EHR) is a core research task in the development of personalized healthcare. Patient EHR data consist of sequences of visits over time, where each visit contains multiple medical codes, including diagnosis, medication, and procedure codes. The most important challenges for this task are to model the temporality and high dimensionality of sequential EHR data and to interpret the prediction results. Existing work solves this problem by employing recurrent neural networks (RNNs) to model EHR data and utilizing simple attention mechanism to interpret the results. However, RNN-based approaches suffer from the problem that the performance of RNNs drops when the length of sequences is large, and the relationships between subsequent visits are ignored by current RNN-based approaches. To address these issues, we propose Dipole, an end-to-end, simple and robust model for predicting patients' future health information. Dipole employs bidirectional recurrent neural networks to remember all the information of both the past visits and the future visits, and it introduces three attention mechanisms to measure the relationships of different visits for the prediction. With the attention mechanisms, Dipole can interpret the prediction results effectively. Dipole also allows us to interpret the learned medical code representations which are confirmed positively by medical experts. Experimental results on two real world EHR datasets show that the proposed Dipole can significantly improve the prediction accuracy compared with the state-of-the-art diagnosis prediction approaches and provide clinically meaningful interpretation.

Drug side-effects become a worldwide public health concern, which are the fourth leading cause of death in the United States. Pharmaceutical industry has paid tremendous effort to identify drug side-effects during the drug development. However, it is impossible and impractical to identify all of them. Fortunately, drug side-effects can also be reported on heterogeneous platforms (i.e., data sources), such as FDA Adverse Event Reporting System and various online communities. However, existing supervised and semi-supervised approaches are not practical as annotating labels are expensive in the medical field. In this paper, we propose a novel and effective unsupervised model Sifter to automatically discover drug side-effects. Sifter enhances the estimation on drug side-effects by learning from various online platforms and measuring platform-level and user-level quality simultaneously. In this way, Sifter demonstrates better performance compared with existing approaches in terms of correctly identifying drug side-effects. Experimental results on five real-world datasets show that Sifter can significantly improve the performance of identifying side-effects compared with the state-of-the-art approaches.

In the age of big data, information for the same entity can be obtained from different sources, which is inevitably conflicting. Therefore, aggregation methods are needed to identify the trustworthy information from such conflicting data. Truth discovery, which improves the aggregation results by estimating source trustworthiness and discovering truths simultaneously, has become an emerging field. Most truth discovery methods assume that sources make their claims independently, which may not be true in practice. As a matter of fact, influences among sources are ubiquitous and the claims made by one source may be influenced by others. Although there is some work that considers source correlation, those methods are designed to handle categorical claims, which is not general enough to represent the complicated real world applications. To tackle these challenges in truth discovery, we propose an unsupervised probabilistic model named IATD. The model takes source correlations as prior for influence derivation. To model influences among sources, we introduce "claim trustworthiness", which fuses the trustworthiness of the source which provides the claim and the trustworthiness of its influencers. Besides, the proposed model can handle different data types using different distributions in the probabilistic model. Experiments on real-world datasets show that IATD model can improve the aggregation performance compared with the state-of-the-art truth discovery approaches. The properties of IATD model are further illustrated using simulated datasets.

Discovering topics in short texts, such as news titles and tweets, has become an important task for many content analysis applications. However, due to the lack of rich context information in short texts, the performance of conventional topic models on short texts is usually unsatisfying. In this paper, we propose a novel topic model for short text corpus using word embeddings. Continuous space word embeddings, which is proven effective at capturing regularities in language, is incorporated into our model to provide additional semantics. Thus we model each short document as a Gaussian topic over word embeddings in the vector space. In addition, considering that background words in a short text are usually not semantically related, we introduce a discrete background mode over word types to complement the continuous Gaussian topics. We evaluate our model on news titles from data sources like abcnews, showing that our model is able to extract more coherent topics from short texts compared with the baseline methods and learn better topic representation for each short document.

The demand for automatic extraction of true information (i.e., truths) from conflicting multi-source data has soared recently. A variety of truth discovery methods have witnessed great successes via jointly estimating source reliability and truths. All existing truth discovery methods focus on providing a point estimator for each object's truth, but in many real-world applications, confidence interval estimation of truths is more desirable, since confidence interval contains richer information. To address this challenge, in this paper, we propose a novel truth discovery method (ETCIBoot) to construct confidence interval estimates as well as identify truths, where the bootstrapping techniques are nicely integrated into the truth discovery procedure. Due to the properties of bootstrapping, the estimators obtained by ETCIBoot are more accurate and robust compared with the state-of-the-art truth discovery approaches. Theoretically, we prove the asymptotical consistency of the confidence interval obtained by ETCIBoot. Experimentally, we demonstrate that ETCIBoot is not only effective in constructing confidence intervals but also able to obtain better truth estimates.

In the past decade, commercial crowdsourcing platforms have revolutionized the ways of classifying and annotating data, especially for large datasets. Obtaining labels for a single instance can be inexpensive, but for large datasets, it is important to allocate budgets wisely. With limited budgets, requesters must trade-off between the quantity of labeled instances and the quality of the final results. Existing budget allocation methods can achieve good quantity but cannot guarantee high quality of individual instances under a tight budget. However, in some scenarios, requesters may be willing to label fewer instances but of higher quality. Moreover, they may have different requirements on quality for different tasks. To address these challenges, we propose a flexible budget allocation framework called Requallo. Requallo allows requesters to set their specific requirements on the labeling quality and maximizes the number of labeled instances that achieve the quality requirement under a tight budget. The budget allocation problem is modeled as a Markov decision process and a sequential labeling policy is produced. The proposed policy greedily searches for the instance to query next as the one that can provide the maximum reward for the goal. The Requallo framework is further extended to consider worker reliability so that the budget can be better allocated. Experiments on two real-world crowdsourcing tasks as well as a simulated task demonstrate that when the budget is tight, the proposed Requallo framework outperforms existing state-of-the-art budget allocation methods from both quantity and quality aspects.

In crowdsourced data aggregation task, there exist conflicts in the answers provided by large numbers of sources on the same set of questions. The most important challenge for this task is to estimate source reliability and select answers that are provided by high-quality sources. Existing work solves this problem by simultaneously estimating sources' reliability and inferring questions' true answers (i.e., the truths). However, these methods assume that a source has the same reliability degree on all the questions, but ignore the fact that sources' reliability may vary significantly among different topics. To capture various expertise levels on different topics, we propose FaitCrowd, a fine grained truth discovery model for the task of aggregating conflicting data collected from multiple users/sources. FaitCrowd jointly models the process of generating question content and sources' provided answers in a probabilistic model to estimate both topical expertise and true answers simultaneously. This leads to a more precise estimation of source reliability. Therefore, FaitCrowd demonstrates better ability to obtain true answers for the questions compared with existing approaches. Experimental results on two real-world datasets show that FaitCrowd can significantly reduce the error rate of aggregation compared with the state-of-the-art multi-source aggregation approaches due to its ability of learning topical expertise from question content and collected answers.