Time series forecasting under distribution shift remains challenging, as existing deep learning models often rely on local statistical normalization (e.g., mean and variance) that fails to capture global distribution shift. Methods like RevIN and its variants attempt to decouple distribution and pattern but still struggle with missing values, noisy observations, and invalid channel-wise affine transformation. To address these limitations, we propose Affine Prototype-Timestamp(APT), a lightweight and flexible plug-in module that injects global distribution features into the normalization–forecasting pipeline. By leveraging timestamp-conditioned prototype learning, APT dynamically generates affine parameters that modulate both input and output series, enabling the backbone to learn from self-supervised, distribution-aware clustered instances. APT is compatible with arbitrary forecasting backbones and normalization strategies while introducing minimal computational overhead. Extensive experiments across six benchmark datasets and multiple backbone-normalization combinations demonstrate that APT significantly improves forecasting performance under distribution shift.

Multi-hop question answering (QA) typically involves retrieving multiple relevant passages as evidence, which a reader then uses to derive the final answer. Existing methods often decompose complex questions into subquestions, leveraging large language models for step-by-step inference. However, these approaches fail to address the interdependence between retrieval and decomposition, resulting in suboptimal performance: Incomplete or inaccurate evidence can lead to poorly crafted subquestions, which, in turn, amplify the retrieval of irrelevant information and impede the reasoning process. To tackle this, we introduce BidLoop, a multi-step reasoning framework that explicitly models Bid irectional Loop between decomposition and retrieval. BidLoop employs four specialized modules: the Planner, Evaluator, Retriever, and Reader. In each reasoning round, the Planner generates a new subquestion based on prior evidence and subquestion-answer pairs. The Evaluator assesses whether the gathered evidence is sufficient to produce the final answer or if further reasoning is needed. Guided by the Planner's subquestion, the Retriever fetches relevant evidence, while the Reader answers the subquestion, adding new evidence for the next round. Our approach excels in generalization, performing strongly on unseen datasets without training on them. Extensive experiments across diverse multi-hop QA datasets demonstrate that BidLoop significantly surpasses existing state-of-the-art models.

Aerial object detection plays a vital role in applications such as natural disaster prevention and urban traffic management, thanks to its ability to handle wide coverage areas and diverse objects. As a leading method for this task, You Only Look Once (YOLO) leverages multi-scale feature extraction to detect objects of various sizes. However, most YOLO-based methods focus on feature extraction and fusion from adjacent scales, neglecting the potential collaboration between non-adjacent scales. This limitation leads to redundant parameters and suboptimal detection performance. To address these issues, this paper proposes AF-YOLO (Asymptotic Feature Extraction and Fusion YOLO), a novel approach tailored for aerial object detection. AF-YOLO introduces two lightweight modules: SCC2f and PAFFN. SCC2f, an optimized version of cross-stage partial bottleneck with spatial and channel reconstruction convolution layers, reduces redundancy and enables efficient multi-scale feature extraction. PAFFN, a parallel asymptotic feature fusion network, facilitates enhanced interaction and fusion of non-adjacent scale features. Additionally, AF-YOLO incorporates a P2 layer to improve small object detection and removes YOLO’s P5 layer for a more lightweight design, specifically optimized for aerial detection tasks. Experimental results demonstrate AF-YOLO’s significant improvements across multiple benchmarks: on the VisDrone dataset, it achieves a 6.1% higher mAP0.5compared to recent baselines while using only 41.8% of their parameters; on the DIOR dataset, it shows a 3.3% accuracy improvement over YOLOv8. These quantitative results are further supported by its superior performance on the DOTA and FAIR1M datasets, with additional validation on HazyDet confirming its robustness in adverse weather conditions. Collectively, these achievements highlight AF-YOLO’s exceptional generalization capability and efficient lightweight design, establishing a new state-of-the-art for aerial object detection systems.

Diffusion models have received wide attention in generation tasks. However, the expensive computation cost prevents the application of diffusion models in resource-constrained scenarios. Quantization emerges as a practical solution that significantly saves storage and computation by reducing the bit-width of parameters. However, the existing quantization methods for diffusion models still cause severe degradation in performance, especially under extremely low bit-widths (2-4 bit). The primary decrease in performance comes from the significant discretization of activation values at low bit quantization. Too few activation candidates are unfriendly for outlier significant weight channel quantization, and the discretized features prevent stable learning over different time steps of the diffusion model. This paper presents MPQ-DM, a Mixed-Precision Quantization method for Diffusion Models. The proposed MPQ-DM mainly relies on two techniques: (1) To mitigate the quantization error caused by outlier severe weight channels, we propose an Outlier-Driven Mixed Quantization (OMQ) technique that uses Kurtosis to quantify outlier salient channels and apply optimized intra-layer mixed-precision bit-width allocation to recover accuracy performance within target efficiency. (2) To robustly learn representations crossing time steps, we construct a Time-Smoothed Relation Distillation (TRD) scheme between the quantized diffusion model and its full-precision counterpart, transferring discrete and continuous latent to a unified relation space to reduce the representation inconsistency. Comprehensive experiments demonstrate that MPQ-DM achieves significant accuracy gains under extremely low bit-widths compared with SOTA quantization methods. MPQ-DM achieves a 58% FID decrease under W2A4 setting compared with baseline, while all other methods even collapse.

In safety-critical domains such as medical diagnostics and autonomous driving, single-image evidence is sometimes insufficient to reflect the inherent ambiguity of vision problems. Therefore, multiple plausible assumptions that match the image semantics may be needed to reflect the actual distribution of targets and support downstream tasks. However, balancing and improving the diversity and consistency of segmentation predictions under the high-dimensional output spaces and potential multimodal distributions is still challenging. This paper presents Hierarchical Self-Regulation Diffusion (HSRDiff), a unified framework that simulates joint probability distribution over entire labels. Our model self-regulates the balance between the two modes of predicting the label and noise in a novel ``differentiation to unification" pipeline and dynamically fits the optimal path to model the aleatoric uncertainty rooted in observations. In addition, we preserve the high-fidelity reconstruction of the delicate structure in images by leveraging the hierarchical multi-scale condition priors. We validate HSRDiff in three different semantic scenarios. Experimental results show that HSRDiff is superior to the comparison method with a considerable performance gap.

Multi-teacher Knowledge Distillation (KD) transfers diverse knowledge from a teacher pool to a student network. The core problem of multi-teacher KD is how to balance distillation strengths among various teachers. Most existing methods often develop weighting strategies from an individual perspective of teacher performance or teacher-student gaps, lacking comprehensive information for guidance. This paper proposes Multi-Teacher Knowledge Distillation with Reinforcement Learning (MTKD-RL) to optimize multi-teacher weights. In this framework, we construct both teacher performance and teacher-student gaps as state information to an agent. The agent outputs the teacher weight and can be updated by the return reward from the student. MTKD-RL reinforces the interaction between the student and teacher using an agent in an RL-based decision mechanism, achieving better matching capability with more meaningful weights. Experimental results on visual recognition tasks, including image classification, object detection, and semantic segmentation tasks, demonstrate that MTKD-RL achieves state-of-the-art performance compared to the existing multi-teacher KD works.

As deep learning models expand, the pre-training-fine-tuning paradigm has become the standard approach for handling various downstream tasks. However, shared parameters can lead to diminished performance when dealing with complex datasets involving multiple tasks. While introducing Mixture-of-Experts (MoE) methods has alleviated this issue to some extent, it also significantly increases the number of parameters required for fine-tuning and training time, introducing greater parameter redundancy. To address these challenges, we propose a method for allocating expert numbers based on parameter sensitivity-LoRA-SMoE (A Sensitivity-Driven Expert Allocation Method in LoRA-MoE for Efficient Fine-Tuning). This method rapidly assesses the sensitivity of different tasks to parameters by sampling a small amount of data and using gradient information. It then adaptively allocates expert numbers within a given budget. The process maintains comparable memory consumption to LoRA (Low-Rank Adaptation) while ensuring an efficient and resource-friendly fine-tuning procedure. Experimental results demonstrate that compared to SOTA fine-tuning methods, our LoRA-SMoE approach can enhance model performance while reducing the number of trainable parameters. This significantly improves model performance in resource-constrained environments. Additionally, due to its efficient parameter sensitivity evaluation mechanism, LoRA-SMoE requires minimal computational overhead to optimize expert allocation, making it particularly suitable for scenarios with limited computational resources. All the code in this study will be made publicly available following the acceptance of the paper for publication. Source code is at https://github.com/EMLS-ICTCAS/LoRA-SMoE

Sparse matrix-vector multiplication (SpMV) is a fundamental operation with a wide range of applications in scientific computing and artificial intelligence. However, the large scale and sparsity of sparse matrix often make it a performance bottleneck. In this paper, we highlight the effectiveness of hash-based techniques in optimizing sparse matrix reordering, introducing the Hash-based Partition (HBP) format, a lightweight SpMV approach. HBP retains the performance benefits of the 2Dpartitioning method while leveraging the hash transformation's ability to group similar elements, thereby accelerating the preprocessing phase of sparse matrix reordering. Additionally, we achieve parallel load balancing across matrix blocks through a competitive method. Our experiments, conducted on both Nvidia Jetson AGX Orin and Nvidia RTX 4090, show that in the preprocessing step, our method offers an average speedup of 3.53 times compared to the sorting approach and 3.67 times compared to the dynamic programming method employed in Regu2D. Furthermore, in SpMV, our method achieves a maximum speedup of 3.32 times on Orin and 3.01 times on RTX4090 against the CSR format in sparse matrices from the University of Florida Sparse Matrix Collection.

Spatio-temporal tasks often encounter incomplete data arising from missing or inaccessible sensors, making spatio-temporal kriging crucial for inferring the completely missing temporal information. However, current models struggle with ensuring the validity and generalizability of inferred spatio-temporal patterns, especially in capturing dynamic spatial dependencies and temporal shifts, and optimizing the generalizability of unknown sensors. To overcome these limitations, we propose Spatio-Temporal Aware Graph Adversarial Neural Network (STA-GANN), a novel GNN-based kriging framework that improves spatio-temporal pattern validity and generalization. STA-GANN integrates (i) Decoupled Phase Module that senses and adjusts for timestamp shifts. (ii) Dynamic Data-Driven Metadata Graph Modeling to update spatial relationships using temporal data and metadata; (iii) An adversarial transfer learning strategy to ensure generalizability. Extensive validation across nine datasets from four fields and theoretical evidence both demonstrate the superior performance of STA-GANN.

Knowledge Graph Embedding (KGE) aims to project entities and relations into a low-dimensional space, so as to enable Knowledge Graphs (KGs) to be effectively used by downstream AI tasks. Most existing KGs (e.g. Wikidata) suffer from the data imbalance issue, i.e., the occurrence frequencies vary significantly among different entities. Current KGE models use a fixed embedding size, leading to overfitting for low-frequency entities and underfitting for high-frequency ones. A simple method is to manually set embedding sizes based on frequency, but this is not feasible due to the complexity and the large number of entities. To this end, we propose CustomizE, which customizes embedding sizes in a data-driven way, assigning larger sizes for high-frequency entities and smaller sizes for low-frequency ones. We use bilevel optimization for stable learning of representations and sizes. It is noteworthy that our framework is universal and flexible, which is suitable for various KGE models. Experiments on link prediction tasks show its superiority over state-of-the-art baselines.

The rapid advancement of diffusion models has increased the need for customized image generation. However, current customization methods face several limitations: 1) typically accept either image or text conditions alone; 2) customization in complex visual scenarios often leads to subject leakage or confusion; 3) image-conditioned outputs tend to suffer from inconsistent backgrounds; and 4) high computational costs. To address these issues, this paper introduces Multi-party Collaborative Attention Control (MCA-Ctrl), a tuning-free method that enables high-quality image customization using both text and complex visual conditions. Specifically, MCA-Ctrl leverages two key operations within the self-attention layer to coordinate multiple parallel diffusion processes and guide the target image generation. This approach allows MCA-Ctrl to capture the content and appearance of specific subjects while maintaining semantic consistency with the conditional input. Additionally, to mitigate subject leakage and confusion issues common in complex visual scenarios, we introduce a Subject Localization Module that extracts precise subject and editable image layers based on user instructions. Extensive quantitative and human evaluation experiments show that MCA-Ctrl outperforms existing methods in zero-shot image customization, effectively resolving the mentioned issues.

Existing Incremental Object Detection (IOD) methods partially alleviate catastrophic forgetting when incrementally detecting new objects in real-world scenarios. However, many of these methods rely on the assumption that unlabeled old-class objects may co-occur with labeled new-class objects in the incremental data. When unlabeled old-class objects are absent, the performance of existing methods tends to degrade. The absence can be mitigated by generating old-class samples, but it incurs high costs. This paper argues that previous generation-based IOD suffers from redundancy, both in the use of generative models, which require additional training and storage, and in the overproduction of generated samples, many of which do not contribute significantly to performance improvements. To eliminate the redundancy, we propose Inversed Objects Replay (IOR). Specifically, we generate old-class samples by inversing the original detectors, thus eliminating the necessity of training and storing additional generative models. We propose augmented replay to reuse the objects in generated samples, reducing redundant generations. Moreover, we propose high-value knowledge distillation focusing on the positions of old-class objects overwhelmed by the background, which transfers the knowledge to the incremental detector. Extensive experiments conducted on MS COCO 2017 demonstrate that our method can efficiently improve detection performance in IOD scenarios with the absence of old-class objects.

Knowledge Distillation (KD) aims to improve a light-weight student network supervised by a large teacher network. The core idea of KD is to explore valuable knowledge from the teacher. Previous works often extract information from a single sample, but ignore relation modeling among multiple samples between student and teacher. Therefore, we propose Cross-Layer Graph Knowledge Distillation (CLGKD) that conducts graph-augmented feature and relation distillation assisted by graph neural networks. We further propose a meta-learning mechanism to optimize cross-layer matching weights for promoting GKD among all student and teacher layers. Experimental results on image classification and object detection demonstrate that CLGKD achieves state-of-the-art performance compared to other KD methods. Our code is available at https://github.com/cynmzzz/ICASSP2025-CLGKD

Open-world object detection (OWOD) is vital for identifying the new objects not encountered during training. Among the various methods for OWOD, Object Proposals without Learning Classification (OPwLC) stands out, with its Object Localization Network (OLN) stressing the localization features. However, OLN overlooks classification features, leading OPwLC to identify parts of a single object as multiple objects mistakenly. Inspired by the non-maximum suppression (NMS) technique, known for eliminating low-confidence detections, we sought to integrate NMS into OPwLC. However, direct integration of NMS into OPwLC presents a challenge, as OLN does not generate classification confidence scores, which are critical for applying NMS. To address this limitation, we developed a confidence measure module and proposed OLN++, filling the confidence scores gap. OLN++ can be easily implemented with just a few fully connected layers. We evaluated the effectiveness of OLN++ using NMS, Soft-NMS, and the Weighted Box Fusion variant on open-world detection tasks. Experimental results demonstrate that OLN++ significantly outperforms the original OLN.

Spatiotemporal forecasting tasks, such as traffic flow, combustion dynamics, and weather forecasting, often require complex models that suffer from low training efficiency and high memory consumption. This paper proposes a lightweight framework, Spectral Decoupled Knowledge Distillation (termed SDKD), which transfers the multi-scale spatiotemporal representations from a complex teacher model to a more efficient lightweight student network. The teacher model follows an encoder-latent evolution-decoder architecture, where its latent evolution module decouples high-frequency details and low-frequency trends using convolution and Transformer (global low-frequency modeler). However, the multi-layer convolution and deconvolution structures result in slow training and high memory usage. To address these issues, we propose a frequency-aligned knowledge distillation strategy, which extracts multi-scale spectral features from the teacher's latent space, including both high and low frequency components, to guide the lightweight student model in capturing both local fine-grained variations and global evolution patterns. Experimental results show that SDKD significantly improves performance, achieving reductions of up to 81.3% in MSE and in MAE 52.3% on the Navier-Stokes equation dataset. The framework effectively captures both high-frequency variations and long-term trends while reducing computational complexity. Our codes are available at https://github.com/itsnotacie/SDKD

Diffusion transformers (DiT) have demonstrated exceptional performance in video generation. However, their large number of parameters and high computational complexity limit their deployment on edge devices. Quantization can reduce storage requirements and accelerate inference by lowering the bit-width of model parameters. Yet, existing quantization methods for image generation models do not generalize well to video generation tasks. We identify two primary challenges: the loss of information during quantization and the misalignment between optimization objectives and the unique requirements of video generation. To address these challenges, we present **Q-VDiT**, a quantization framework specifically designed for video DiT models. From the quantization perspective, we propose the *Token aware Quantization Estimator* (TQE), which compensates for quantization errors in both the token and feature dimensions. From the optimization perspective, we introduce *Temporal Maintenance Distillation* (TMD), which preserves the spatiotemporal correlations between frames and enables the optimization of each frame with respect to the overall video context. Our W3A6 Q-VDiT achieves a scene consistency score of 23.40, setting a new benchmark and outperforming the current state-of-the-art quantization methods by **1.9$\times$**.

3D few-shot class incremental learning (FSCIL) aims to learn new point cloud categories from limited samples while preventing the forgetting of previously learned categories. This research area significantly enhances the capabilities of self-driving vehicles and computer vision systems. Existing 3D FSCIL approaches primarily utilize multimodal pre-trained models to extract the semantic features, heavily dependent on meticulously designed high-quality prompts and fine-tuning strategies. To reduce this dependence, this paper proposes a novel method for **3D** **F**SCI**L** with **E**mbedded **G**eometric features (**3D-FLEG**). Specifically, 3D-FLEG develops a point cloud *geometric feature extraction module* to capture category-related geometric characteristics. To address the modality heterogeneity issues that arise from integrating geometric and text features, 3D-FLEG introduces a *geometric feature embedding module*. By augmenting text prompts with spatial geometric features through these modules, 3D-FLEG can learn robust representations of new categories even with limited samples, while mitigating forgetting of the previously learned categories. Experiments conducted on several publicly available 3D point cloud datasets, including ModelNet, ShapeNet, ScanObjectNN, and CO3D, demonstrate 3D-FLEG's superiority over existing state-of-the-art 3D FSCIL methods. Code is available at https://github.com/lixiangqi707/3D-FLEG.

High-performance deep learning depends on efficient tensor programs. In recent years, automatic tensor program optimization, also known as tensor compilation, has emerged as the primary approach to generating efficient tensor programs. However, how to generate kernels with higher performance in a shorter time is still the key challenge. In this paper, we present Gensor, a graph-based construction tensor compilation method for deep learning, to further improve the performance of construction tensor compilation. Unlike existing tree-based methods, Gensor abstracts construction space into a graph structure. Gensor then explores the construction space with Markov analysis. Gensor takes tensor programs as states and models scheduling primitives as transition actions between these states. Therefore, the process of tensor program construction optimization is abstracted as a graph traversal process. This approach expands the optimization space, improving operator performance while ensuring rapid optimization. Extensive experiments with typical operators demonstrate that Gensor significantly outperforms the state-of-the-art methods on GPUs for both cloud servers and edge devices. As a result, Gensor can generate operator kernels in seconds, with performance increasing by 18 % on average, reaching a maximum of 30 %. It also achieves high speedup for end-to-end models like ResNet50 and GPT-2, with an average acceleration of 20 %.

False alarm suppression is becoming increasingly important as it directly impacts the reliability and efficiency of synthetic aperture radar (SAR) image detection systems. previous methods for false alarm suppression have focused primarily on identifying the motion properties of targets and removing the embedded noise. However, SAR images are captured in a single band, which means they lack continuous bands and dynamic information. In addition, noise removal often results in a significant loss of detail in the image. In this paper, we innovatively propose a classification-based false alarm suppression framework for SAR object detection, avoiding the need for motion identification and noise removal. In practice, we first train a classification network to categorize the SAR image slices into ocean, land, and offshore scenes. Based on the classification results, we then dynamically adjust the Intersection Over Union (IoU) threshold of Non-Maximum Suppression (NMS) in different scenes. Experimental results on a newly large multi-class target SAR dataset, MSAR-1.0, show that the false alarm rate decreased from 21% to 13%.

The advent of universal time series forecasting models has revolutionized zero-shot forecasting across diverse domains, yet the critical role of data diversity in training these models remains underexplored. Existing large-scale time series datasets often suffer from inherent biases and imbalanced distributions, leading to suboptimal model performance and generalization. To address this gap, we introduce BLAST, a novel pre-training corpus designed to enhance data diversity through a balanced sampling strategy. First, BLAST incorporates 321 billion observations from publicly available datasets and employs a comprehensive suite of statistical metrics to characterize time series patterns. Then, to facilitate pattern-oriented sampling, the data is implicitly clustered using grid-based partitioning. Furthermore, by integrating grid sampling and grid mixup techniques, BLAST ensures a balanced and representative coverage of diverse patterns. Experimental results demonstrate that models pre-trained on BLAST achieve state-of-the-art performance with a fraction of the computational resources and training tokens required by existing methods. Our findings highlight the pivotal role of data diversity in improving both training efficiency and model performance for the universal forecasting task.

Multivariate Time Series Forecasting (MTSF) involves predicting future values of multiple interrelated time series. Recently, deep learning-based MTSF models have gained significant attention for their promising ability to mine semantics (global and local information) within MTS data. However, these models are pervasively susceptible to missing values caused by malfunctioning data collectors. These missing values not only disrupt the semantics of MTS, but their distribution also changes over time. Nevertheless, existing models lack robustness to such issues, leading to suboptimal forecasting performance. To this end, in this paper, we propose Multi-View Representation Learning (Merlin), which can help existing models achieve semantic alignment between incomplete observations with different missing rates and complete observations in MTS. Specifically, Merlin consists of two key modules: offline knowledge distillation and multi-view contrastive learning. The former utilizes a teacher model to guide a student model in mining semantics from incomplete observations, similar to those obtainable from complete observations. The latter improves the student model's robustness by learning from positive/negative data pairs constructed from incomplete observations with different missing rates, ensuring semantic alignment across different missing rates. Therefore, Merlin is capable of effectively enhancing the robustness of existing models against unfixed missing rates while preserving forecasting accuracy. Experiments on four real-world datasets demonstrate the superiority of Merlin.

Benefiting from high capacity for capturing complex temporal patterns, deep learning (DL) has significantly advanced time series forecasting (TSF). However, deep models tend to suffer from severe overfitting due to the inherent vulnerability of time series to noise and anomalies. The prevailing DL paradigm uniformly optimizes all timesteps through the MSE loss and learns those uncertain and anomalous timesteps without difference, ultimately resulting in overfitting. To address this, we propose a novel selective learning strategy for deep TSF. Specifically, selective learning screens a subset of the whole timesteps to calculate the MSE loss in optimization, guiding the model to focus on generalizable timesteps while disregarding non-generalizable ones. Our framework introduces a dual-mask mechanism to target timesteps: (1) an uncertainty mask leveraging residual entropy to filter uncertain timesteps, and (2) an anomaly mask employing residual lower bound estimation to exclude anomalous timesteps. Extensive experiments across eight real-world datasets demonstrate that selective learning can significantly improve the predictive performance for typical state-of-the-art deep models, including 37.4% MSE reduction for Informer, 8.4% for TimesNet, and 6.5% for iTransformer.

Transformers have gained attention in atmospheric time series forecasting (ATSF) for their ability to capture global spatial-temporal correlations. However, their complex architectures lead to excessive parameter counts and extended training times, limiting their scalability to large-scale forecasting. In this paper, we revisit ATSF from a theoretical perspective of atmospheric dynamics and uncover a key insight: spatial-temporal position embedding (STPE) can inherently model spatial-temporal correlations even without attention mechanisms. Its effectiveness arises from integrating geographical coordinates and temporal features, which are intrinsically linked to atmospheric dynamics. Based on this, we propose **STELLA**, a **S**patial-**T**emporal knowledge **E**mbedded **L**ightweight mode**L** for ASTF, utilizing only STPE and an MLP architecture in place of Transformer layers. With 10k parameters and one hour of training, STELLA achieves superior performance on five datasets compared to other advanced methods. The paper emphasizes the effectiveness of spatial-temporal knowledge integration over complex architectures, providing novel insights for ATSF.

Spatio-temporal trajectory representation learning plays a crucial role in various urban applications such as transportation systems, urban planning, and environmental monitoring. Existing methods can be divided into single-view and multi-view approaches, with the latter offering richer representations by integrating multiple sources of spatio-temporal data. However, these methods often struggle to generalize across diverse urban scenes due to multi-city structural heterogeneity, which arises from the disparities in road networks, grid layouts, and traffic regulations across cities, and the amplified seesaw phenomenon, where optimizing for one city, view, or task can degrade performance in others. These challenges hinder the deployment of trajectory learning models across multiple cities, limiting their real-world applicability. In this work, we propose SMARTraj$^2$, a novel stable multi-city adaptive method for multi-view spatio-temporal trajectory representation learning. Specifically, we introduce a feature disentanglement module to separate domain-invariant and domain-specific features, and a personalized gating mechanism to dynamically stabilize the contributions of different views and tasks. Our approach achieves superior generalization across heterogeneous urban scenes while maintaining robust performance across multiple downstream tasks. Extensive experiments on benchmark datasets demonstrate the effectiveness of SMARTraj$^2$ in enhancing cross-city generalization and outperforming state-of-the-art methods. See our project website at \url{https://github.com/GestaltCogTeam/SMARTraj}.

Abstract GEMM (General Matrix Multiplication) serves as a fundamental operator for deep learning computations. Especially in attention-based deep learning models, such as Bert, GPT, and SAM, the sizes of matrices involved in GEMMs exhibit an unbalanced distribution due to the variable input, resulting in the low utilization of hardware resources. To address the issue, this paper proposes inserting a novel GEMM processing layer into the deep learning inference stack and using an adaptive load balancing method to partition and schedule GEMM computation tasks. The method is implemented with hardware runtime resource information, such as the occupancy of computing units, etc. Experiment results show the remarkable performance of our method in unbalanced input GEMM scenarios, achieving an average performance improvement of 2.3x. The method also performs well in attention-based models (GPT-2 and SAM), achieving an average inference speed improvement of 1.1x. These findings highlight the effectiveness of resource-aware algorithm optimization, especially for computation task scheduling.

Knowledge Graph Embedding (KGE) aims to learn dense embeddings as the representations for entities and relations in KGs. Indeed, the entities in existing KGs suffer from the data imbalance issue, i.e., there exists a substantial disparity in the occurrence frequencies among various entities. Existing KGE models pre-define a unified and fixed dimension size for all entity embeddings. However, embedding sizes of entities are highly desired for their frequencies, while a uniform embedding size may result in inadequate expression of entities, i.e., leading to overfitting for low-frequency entities and underfitting for high-frequency ones. A straight-forward idea is to set the embedding sizes for each entity before KGE training. However, manually selecting different embedding sizes is labor-intensive and time-consuming, which is difficult to achieve in real-world scenarios. To tackle this problem, we propose AdaE, which adaptively learns KG embeddings with different embedding sizes during training. In particular, AdaE is capable of selecting appropriate dimension sizes for each entity from a continuous integer space. To this end, we specially tailor bilevel optimization for the KGE task, which alternately learns representations and embedding sizes of entities. Moreover, it is worth noting that our framework is general and flexible, which is suitable for various existing KGE models. Extensive experiments demonstrate the effectiveness and compatibility of AdaE.

Multivariate Time Series (MTS) analysis is crucial to understanding and managing complex systems, such as traffic and energy systems, and a variety of approaches to MTS forecasting have been proposed recently. However, we often observe inconsistent or seemingly contradictory performance findings across different studies. This hinders our understanding of the merits of different approaches and slows down progress. We address the need for means of assessing MTS forecasting proposals reliably and fairly, in turn enabling better exploitation of MTS as seen in different applications. Specifically, we first propose BasicTS+, a benchmark designed to enable fair, comprehensive, and reproducible comparison of MTS forecasting solutions. BasicTS+ establishes a unified training pipeline and reasonable settings, enabling an unbiased evaluation. Second, we identify the heterogeneity across different MTS as an important consideration and enable classification of MTS based on their temporal and spatial characteristics. Disregarding this heterogeneity is a prime reason for difficulties in selecting the most promising technical directions. Third, we apply BasicTS+ along with rich datasets to assess the capabilities of more than 30 MTS forecasting solutions. This provides readers with an overall picture of the cutting-edge research on MTS forecasting.

Spatial-Temporal Graph Neural Networks (STGNNs) have been widely utilized in multivariate time series forecasting (MTSF), but they rely on the assumption of data completeness. In practice, due to factors such as natural disaster, STGNNs frequently encounter the challenge of missing data resulting from numerous malfunctioning data collectors. In this case, on the one hand, due to the presence of missing values, STGNNs easily generate incorrect spatial correlations, leading to the performance degradation. On the other hand, STGNNs require separate training of models for different missing rates, limiting their robustness. To address these challenges, we first propose two important components (interpolation attention and adaptive graph convolution), which utilize normal values to recover missing values into reliable representations and reconstruct spatial correlations. Then, we replace the fully connected layers in simple recursive units with these two components and propose Graph Interpolation Attention Recursive Network (GinAR), aiming to recursively correct spatial correlations and achieve end-to-end MTSF with missing values. Finally, we use data with different missing rates as positive and negative data pairs. By employing contrastive learning to train GinAR, we propose GinAR+ and enhance its robustness to data with different missing rates. Experiments validate the superiority of GinAR+ and our motivation.

Classical sequential recommendation models generally adopt ID embeddings to store knowledge learned from user historical behaviors and represent items. However, these unique IDs are challenging to be transferred to new domains. With the thriving of pre-trained language model (PLM), some pioneer works adopt PLM for pre-trained recommendation, where modality information is considered universal across domains via PLM. Unfortunately, the behavioral information in ID embeddings is verified to currently dominate in recommendation compared to modality information and thus limits these models’ performance. In this work, we propose a novel ID-centric recommendation pre-training paradigm (IDP), which directly transfers informative ID embeddings learned in pre-training domains to item representations in new domains. Specifically, in pre-training stage, besides the ID-based sequential recommendation model, we also build a Cross-domain ID-matcher (CDIM) learned by both behavioral and modality information. In the tuning stage, modality information of new domain items is regarded as a cross-domain bridge built by CDIM. They first adopted to retrieve behaviorally and semantically similar items from pre-training domains using CDIM. Next, these retrieved items’ pre-trained ID embeddings are directly adopted to generate downstream new items’ embeddings. Through extensive experiments on real-world datasets, we demonstrate that our proposed model significantly outperforms all baselines.

Class incremental learning (CIL) aims to solve the notorious forgetting problem, which refers to the fact that once the network is updated on a new task, its performance on previously-learned tasks degenerates catastrophically. Most successful CIL methods store exemplars (samples of learned tasks) to train a feature extractor incrementally, or store prototypes (features of learned tasks) to estimate the incremental feature distribution. However, the stored exemplars would violate the data privacy concerns, while the fixed prototypes might not reasonably be consistent with the incremental feature distribution, hindering the exploration of real-world CIL applications. In this paper, we propose a data-free CIL method with embedding distillation and Task-oriented generation (eTag), which requires neither exemplar nor prototype. Embedding distillation prevents the feature extractor from forgetting by distilling the outputs from the networks' intermediate blocks. Task-oriented generation enables a lightweight generator to produce dynamic features, fitting the needs of the top incremental classifier. Experimental results confirm that the proposed eTag considerably outperforms state-of-the-art methods on several benchmark datasets.

Temporal Knowledge Graph Question Answering (TKGQA) aims to answer questions with temporal intent over Temporal Knowledge Graphs (TKGs). The core challenge of this task lies in understanding the complex semantic information regarding multiple types of time constraints (e.g., before, first) in questions. Existing end-to-end methods implicitly model the time constraints by learning time-aware embeddings of questions and candidate answers, which is far from understanding the question comprehensively. Motivated by semantic-parsing-based approaches that explicitly model constraints in questions by generating logical forms with symbolic operators, we design fundamental temporal operators for time constraints and introduce a novel self-improvement Programming method for TKGQA (Prog-TQA). Specifically, Prog-TQA leverages the in-context learning ability of Large Language Models (LLMs) to understand the combinatory time constraints in the questions and generate corresponding program drafts with a few examples given. Then, it aligns these drafts to TKGs with the linking module and subsequently executes them to generate the answers. To enhance the ability to understand questions, Prog-TQA is further equipped with a self-improvement strategy to effectively bootstrap LLMs using high-quality self-generated drafts. Extensive experiments demonstrate the superiority of the proposed Prog-TQA on MultiTQ and CronQuestions datasets, especially in the Hits@1 metric.

We propose a novel regularization method to effectively train a neural network for avoiding overfitting, thus improving the performance. The core idea is to bridge the gap between predictive distributions derived from two popular image mixture techniques Mixup and CutMix by an ensemble distribution in a class-wise manner. Consistent optimization towards these three distributions is conducted by mutual distillation to guide the model to alleviate over-confidence predictions and robustly learn discriminative features as the classification evidence. Experiments across various image classification tasks show that our method significantly achieves better performance than previous data augmentation Mixup+CutMix and Self-KD methods.

Existing online Knowledge Distillation (KD) often perform probability-based predictions from independent data samples for knowledge transfer. However, these online KD methods neglect valuable relational information across multiple networks. To address this problem, we propose Online Relational Knowledge Distillation (ORKD). ORKD includes a discriminative loss to construct meaningful feature space and a relational distillation loss to guide structured knowledge transfer among multiple networks. Beyond feature-level distillation, we further construct an ensemble teacher by aggregating probability predictions from multiple networks. The virtual teacher is used to supervise a specific network to enhance its accuracy and avoid the cohort homogenization problem. Experimental results on CIFAR-100 and ImageNet classification demonstrate that ORKD achieves the best performance among state-of-the-art online KD methods over various network architectures. The qualitative visualization shows that ORKD can help the network to learn a more discriminative feature space, resulting in better classification performance.

Contrastive Language-Image Pre-training (CLIP) has become a promising language-supervised visual pre-training framework. This paper aims to distill small CLIP models supervised by a large teacher CLIP model. We propose several distillation strategies, including relation, feature, gradient and contrastive paradigms, to examine the effectiveness of CLIP-Knowledge Distillation (KD). We show that a simple feature mimicry with Mean Squared Error loss works surprisingly well. Moreover, interactive contrastive learning across teacher and student encoders is also effective in performance improvement. We explain that the success of CLIP-KD can be attributed to maximizing the feature similarity between teacher and student. The unified method is applied to distill several student models trained on CC3M+12M. CLIP-KD improves student CLIP models consistently over zero-shot ImageNet classification and cross-modal retrieval bench-marks. When using ViT-U14 pretrained on Laion-400M as the teacher, CLIP-KD achieves 57.5% and 55.4% zero-shot top-1 ImageNet accuracy over ViT-B/16 and ResNet-50, surpassing the original CLIP without KD by 20.5% and 20.1% margins, respectively. Our code is released on https://github.com/winycg/CLIP-KD.

Complex spatial dependencies in transportation networks make traffic prediction extremely challenging. Much existing work is devoted to learning dynamic graph structures among sensors, and the strategy of mining spatial dependencies from traffic data, known as data-driven, tends to be an intuitive and effective approach. However, Time-Shift of traffic patterns and noise induced by random factors hinder data-driven spatial dependence modeling. In this paper, we propose a novel dynamic frequency domain graph convolution network (DFDGCN) to capture spatial dependencies. Specifically, we mitigate the effects of time-shift by Fourier transform, and introduce the identity embedding of sensors and time embedding when capturing data for graph learning since traffic data with noise is not entirely reliable. The graph is combined with static predefined and self-adaptive graphs during graph convolution to predict future traffic data through classical causal convolutions. Extensive experiments on four real-world datasets demonstrate that our model is effective and outperforms the baselines.

Multi-agent trajectory prediction, as a critical task in modeling complex interactions of objects in dynamic systems, has attracted significant research attention in recent years. Despite the promising advances, existing studies all follow the assumption that data distribution observed during model learning matches that encountered in real-world deployments. However, this assumption often does not hold in practice, as inherent distribution shifts might exist in the mobility patterns for deploy-ment environments, thus leading to poor domain generalization and performance degradation. Consequently, it is appealing to leverage trajectories from multiple source domains to mitigate such discrepancies for multi-agent trajectory prediction task. However, the development of multi-source domain generalization in this task presents two notable issues: (1) negative transfer; (2) inadequate modeling for external factors. To address these issues, we propose a new causal formulation to explicitly model four types of features: domain-invariant and domain-specific features for both the focal agent and neighboring agents. Building upon the new formulation, we propose AdapTraj, a multi-source domain generalization framework specifically tailored for multi-agent trajectory prediction. AdapTraj serves as a plug-and-play module that is adaptable to a variety of models. Extensive experiments on four datasets with different domains demonstrate that AdapTraj consistently outperforms other baselines by a substantial margin.

Policy Distillation (PD) has become an effective method to improve deep reinforcement learning tasks. The core idea of PD is to distill policy knowledge from a teacher agent to a student agent. However, the teacher-student framework requires a well-trained teacher model which is computationally expensive. In the light of online knowledge distillation, we study the knowledge transfer between different policies that can learn diverse knowledge from the same environment. In this work, we propose Online Policy Distillation (OPD) with Decision-Attention (DA), an online learning framework in which different policies operate in the same environment to learn different perspectives of the environment and transfer knowledge to each other to obtain better performance together. With the absence of a well-performance teacher policy, the group-derived targets play a key role in transferring group knowledge to each student policy. However, naive aggregation functions tend to cause student policies quickly homogenize. To address the challenge, we introduce the Decision-Attention module to the online policies distillation framework. The Decision-Attention module can generate a distinct set of weights for each policy to measure the importance of group members. We use the Atari platform for experiments with various reinforcement learning algorithms, including PPO and DQN. In different tasks, our method can perform better than an independent training policy on both PPO and DQN algorithms. This suggests that our OPD-DA can transfer knowledge between different policies well and help agents obtain more rewards.

Online games often house virtual shops where players can acquire character skins. Our task is centered on tailoring skin recommendations across diverse scenarios by analyzing historical interactions such as clicks, usage, and purchases. Traditional multi-behavior recommendation models employed for this task are limited. They either only predict skins based on a single type of behavior or merely recommend skins for target behavior type/task. These models lack the ability to control predictions of skins that are associated with different scenarios and behaviors. To overcome these limitations, we utilize the pretraining capabilities of Large Sequential Models (LSMs) coupled with a novel stimulus prompt mechanism and build a controllable multi-behavior recommendation (CMBR) model. In our approach, the pretraining ability is used to encapsulate users' multi-behavioral sequences into the representation of users' general interests. Subsequently, our designed stimulus prompt mechanism stimulates the model to extract scenario-related interests, thus generating potential skin purchases (or clicks and other interactions) for users. To the best of our knowledge, this is the first work to provide controlled multi-behavior recommendations, and also the first to apply the pretraining capabilities of LSMs in game domain. Through offline experiments and online A/B tests, we validate our method significantly outperforms baseline models, exhibiting about a tenfold improvement on various metrics during the offline test.

The graph-based recommendation has achieved great success in recent years. However, most existing graph-based recommendations focus on capturing user preference based on positive edges/feedback, while ignoring negative edges/feedback (e.g., dislike, low rating) that widely exist in real-world recommender systems. How to utilize negative feedback in graph-based recommendations still remains underexplored. In this study, we first conducted a comprehensive experimental analysis and found that (1) existing graph neural networks are not well-suited for modeling negative feedback, which acts as a high-frequency signal in a user-item graph. (2) The graph-based recommendation suffers from the representation degeneration problem. Based on the two observations, we propose a novel model that models positive and negative feedback from a frequency filter perspective called Dual-frequency Graph Neural Network for Sign-aware Recommendation (DFGNN). Specifically, in DFGNN, the designed dual-frequency graph filter (DGF) captures both low-frequency and high-frequency signals that contain positive and negative feedback. Furthermore, the proposed signed graph regularization is applied to maintain the user/item embedding uniform in the embedding space to alleviate the representation degeneration problem. Additionally, we conduct extensive experiments on real-world datasets and demonstrate the effectiveness of the proposed model. Codes of our model will be released upon acceptance.

Multivariate time series forecasting (MTSF) is crucial for decision-making to precisely forecast the future values/trends, based on the complex relationships identified from historical observations of multiple sequences. Recently, Spatial-Temporal Graph Neural Networks (STGNNs) have gradually become the theme of MTSF model as their powerful capability in mining spatial-temporal dependencies, but almost of them heavily rely on the assumption of historical data integrity. In reality, due to factors such as data collector failures and time-consuming repairment, it is extremely challenging to collect the whole historical observations without missing any variable. In this case, STGNNs can only utilize a subset of normal variables and easily suffer from the incorrect spatial-temporal dependency modeling issue, resulting in the degradation of their forecasting performance. To address the problem, in this paper, we propose a novel Graph Interpolation Attention Recursive Network (named GinAR) to precisely model the spatial-temporal dependencies over the limited collected data for forecasting. In GinAR, it consists of two key components, that is, interpolation attention and adaptive graph convolution to take place of the fully connected layer of simple recursive units, and thus are capable of recovering all missing variables and reconstructing the correct spatial-temporal dependencies for recursively modeling of multivariate time series data, respectively. Extensive experiments conducted on five real-world datasets demonstrate that GinAR outperforms 11 SOTA baselines, and even when 90% of variables are missing, it can still accurately predict the future values of all variables.

Recommendation systems, which assist users in discovering their preferred items among numerous options, have served billions of users across various online platforms. Intuitively, users' interactions with items are highly driven by their unchanging inherent intents (e.g., always preferring high-quality items) and changing demand intents (e.g., wanting a T-shirt in summer but a down jacket in winter). However, both types of intents are implicitly expressed in recommendation scenario, posing challenges in leveraging them for accurate intent-aware recommendations. Fortunately, in search scenario, often found alongside recommendation on the same online platform, users express their demand intents explicitly through their query words. Intuitively, in both scenarios, a user shares the same inherent intent and the interactions may be influenced by the same demand intent. It is therefore feasible to utilize the interaction data from both scenarios to reinforce the dual intents for joint intent-aware modeling. But the joint modeling should deal with two problems: 1) accurately modeling users' implicit demand intents in recommendation; 2) modeling the relation between the dual intents and the interactive items. To address these problems, we propose a novel model named Unified Dual-Intents Translation for joint modeling of Search and Recommendation (UDITSR). To accurately simulate users' demand intents in recommendation, we utilize real queries from search data as supervision information to guide its generation. To explicitly model the relation among the triplet , we propose a dual-intent translation propagation mechanism to learn the triplet in the same semantic space via embedding translations. Extensive experiments demonstrate that UDITSR outperforms SOTA baselines both in search and recommendation tasks.

Continual learning (CL) is designed to learn new tasks while preserving existing knowledge. Replaying samples from earlier tasks has proven to be an effective method to mitigate the forgetting of previously acquired knowledge. However, the current research on the training efficiency of rehearsal-based methods is insufficient, which limits the practical application of CL systems in resource-limited scenarios. The human visual system (HVS) exhibits varying sensitivities to different frequency components, enabling the efficient elimination of visually redundant information. Inspired by HVS, we propose a novel framework called Continual Learning in the Frequency Domain (CLFD). To our knowledge, this is the first study to utilize frequency domain features to enhance the performance and efficiency of CL training on edge devices. For the input features of the feature extractor, CLFD employs wavelet transform to map the original input image into the frequency domain, thereby effectively reducing the size of input feature maps. Regarding the output features of the feature extractor, CLFD selectively utilizes output features for distinct classes for classification, thereby balancing the reusability and interference of output features based on the frequency domain similarity of the classes across various tasks. Optimizing only the input and output features of the feature extractor allows for seamless integration of CLFD with various rehearsal-based methods. Extensive experiments conducted in both cloud and edge environments demonstrate that CLFD consistently improves the performance of state-of-the-art (SOTA) methods in both precision and training efficiency. Specifically, CLFD can increase the accuracy of the SOTA CL method by up to 6.83% and reduce the training time by 2.6×.

Knowledge distillation (KD) is an effective framework that aims to transfer meaningful information from a large teacher to a smaller student. Generally, KD often involves how to define and transfer knowledge. Previous KD methods often focus on mining various forms of knowledge, for example, feature maps and refined information. However, the knowledge is derived from the primary supervised task, and thus, is highly task-specific. Motivated by the recent success of self-supervised representation learning, we propose an auxiliary self-supervision augmented task to guide networks to learn more meaningful features. Therefore, we can derive soft self-supervision augmented distributions as richer dark knowledge from this task for KD. Unlike previous knowledge, this distribution encodes joint knowledge from supervised and self-supervised feature learning. Beyond knowledge exploration, we propose to append several auxiliary branches at various hidden layers, to fully take advantage of hierarchical feature maps. Each auxiliary branch is guided to learn self-supervision augmented tasks and distill this distribution from teacher to student. Overall, we call our KD method a hierarchical self-supervision augmented KD (HSSAKD). Experiments on standard image classification show that both offline and online HSSAKD achieves state-of-the-art performance in the field of KD. Further transfer experiments on object detection further verify that HSSAKD can guide the network to learn better features. The code is available at https://github.com/winycg/HSAKD.

Recently, with the aim of enhancing the trustworthiness of recommender systems, explainable recommendation has attracted much attention from the research community. Intuitively, users’ opinions toward different aspects of an item determine their ratings (i.e., users’ preferences) for the item. Therefore, rating prediction from the perspective of opinions can realize personalized explanations at the level of item aspects and user preferences. However, there are several challenges in developing an opinion-based explainable recommendation: (1) The complicated relationship between users’ opinions and ratings. (2) The difficulty of predicting the potential (i.e., unseen) user-item opinions because of the sparsity of opinion information. To tackle these challenges, we propose an overall preference-aware opinion-based explainable rating prediction model by jointly modeling the multiple observations of user-item interaction (i.e., review, opinion, rating). To alleviate the sparsity problem and raise the effectiveness of opinion prediction, we further propose a triple dual learning-based framework with a novelly designed triple dual constraint . Finally, experiments on three popular datasets show the effectiveness and great explanation performance of our framework.

Real-time applications that require timely data delivery over wireless multi-hop networks within specified deadlines are growing increasingly. Effective routing protocols that can guarantee real-time QoS are crucial, yet challenging, due to the unpredictable variations in end-to-end delay caused by unreliable wireless channels. In such conditions, the upper bound on the end-to-end delay, i.e., worst-case end-to-end delay, should be guaranteed within the deadline. However, existing routing protocols with guaranteed delay bounds cannot strictly guarantee real-time QoS because they assume that the worst-case end-to-end delay is known and ignore the impact of routing policies on the worst-case end-to-end delay determination. In this paper, we relax this assumption and propose DDRL-ARGB, an Adaptive Routing with Guaranteed delay Bounds using Deep Distributional Reinforcement Learning (DDRL). DDRL-ARGB adopts DDRL to jointly determine the worst-case end-to-end delay and learn routing policies. To accurately determine worst-case end-to-end delay, DDRL-ARGB employs a quantile regression deep Q-network to learn the end-to-end delay cumulative distribution. To guarantee real-time QoS, DDRL-ARGB optimizes routing decisions under the constraint of worst-case end-to-end delay within the deadline. To improve traffic congestion, DDRL-ARGB considers the network congestion status when making routing decisions. Extensive results show that DDRL-ARGB can accurately calculate worst-case end-to-end delay, and can strictly guarantee real-time QoS under a small tolerant violation probability against two state-of-the-art routing protocols.

This paper addresses the problem of scheduling real-time wireless flows with general traffic patterns in dynamic network conditions. The main goal is to maximize the fraction of packets to be delivered within their deadlines, which is referred to as timely-throughput. While scheduling algorithms for frame-based traffic models and greedy maximal scheduling methods like LDF have been thoroughly studied, algorithms providing deadline guarantees on packet delivery for general traffic under dynamic network conditions are insufficient. To address this issue, we present a comparative study of two deep reinforcement learning-based scheduling algorithms: RL-Centralized and RL-Decentralized, which are designed to optimize timely-throughput for real-time wireless flows with general traffic patterns in dynamic wireless networks. The RL-Centralized scheduling algorithm formulates the centralized scheduling problem as a Markov Decision Process (MDP) and leverages a Multi-Environments Dueling Double Deep Q-Network (ME-D3QN) structure to adapt to dynamic network conditions. The RL-Decentralized scheduling problem is formulated as a Multi-Agent Markov Decision Process (MMDP) and employs the Node State Consensus Protocol (NSCP) and Lifelong Reinforcement Learning Decentralized Training and Decentralized Execution (LRL-DTDE) structure to accelerate training. Our experimental results indicate that both proposed algorithms can converge quickly and efficiently adapt to dynamic network conditions with better performance than their baseline policies. Finally, test-bed experiments validate simulation results and confirm that the proposed algorithms are practical on resource-limited platforms.=-1

Knowledge graphs (KGs) play a pivotal role in AI-related applications.In order to construct or continuously enrich KGs, automatic knowledge construction and update mechanisms are usually utilized, which inevitably bring in plenty of noise, and noise would degrade the performance of downstream applications.Existing KG error detection methods utilize the embeddings of entities and relations, or directly leverage the paths between entities to test the plausibility of triples, while ignore the valuable hierarchical information contained in the paths between entities.Indeed, the paths between a pair of entities conform to a hierarchical structure.Specifically, there may be a number of paths between two entities, and each path is comprised of several relations.The hierarchical structure is able to provide precious information, and is beneficial to leverage the path information in a fine-grained manner.To this end, in this paper, we propose a novel model named KG error detection with HiErarchical pAth stRucture (HEAR for short).Particularly, for a given triple, HEAR first learns path representations with the relations contained in the path, then integrates all path representations, and at last predicts the plausibility of the triple.Finally, we extensively validate the superiority of HEAR against various state-of-the-art baselines.

Large-scale Multivariate Time Series(MTS) widely exist in various real-world systems, imposing significant demands on model efficiency. A recent work, STID, addressed the high complexity issue of popular Spatial-Temporal Graph Neural Networks(STGNNs). Despite its success, when applied to large-scale MTS data, the number of parameters of STID for modeling spatial dependencies increases substantially, leading to over-parameterization issues and suboptimal performance. These observations motivate us to explore new approaches for modeling spatial dependencies in a parameter-friendly manner. In this paper, we argue that the spatial properties of variables are essentially the superposition of multiple cluster centers. Accordingly, we propose a Cluster-Aware Network(CANet), which effectively captures spatial dependencies by mining the implicit cluster centers of variables. CANet solely optimizes the cluster centers instead of the spatial information of all nodes, thereby significantly reducing the parameter amount. Extensive experiments on two large-scale datasets validate our motivation and demonstrate the superiority of CANet.

Multivariate time series long-term prediction, which aims to predict the change of data in a long time, can provide references for decision-making. Although transformer-based models have made progress in this field, they usually do not make full use of three features of multivariate time series: global information, local information, and variables correlation. To effectively mine the above three features and establish a high-precision prediction model, we propose a double sampling transformer (DSformer), which consists of the double sampling (DS) block and the temporal variable attention (TVA) block. Firstly, the DS block employs down sampling and piecewise sampling to transform the original series into feature vectors that focus on global information and local information respectively. Then, TVA block uses temporal attention and variable attention to mine these feature vectors from different dimensions and extract key information. Finally, based on a parallel structure, DSformer uses multiple TVA blocks to mine and integrate different features obtained from DS blocks respectively. The integrated feature information is passed to the generative decoder based on a multi-layer perceptron to realize multivariate time series long-term prediction. Experimental results on nine real-world datasets show that DSformer can outperform eight existing baselines.

Takeaway recommender systems, which aim to accurately provide stores that offer foods meeting users' interests, have served billions of users in our daily life. Different from traditional recommendation, takeaway recommendation faces two main challenges: (1) Dual Interaction-Aware Preference Modeling. Traditional recommendation commonly focuses on users' single preferences for items while takeaway recommendation needs to comprehensively consider users' dual preferences for stores and foods. (2) Period-Varying Preference Modeling. Conventional recommendation generally models continuous changes in users' preferences from a session-level or day-level perspective. However, in practical takeaway systems, users' preferences vary significantly during the morning, noon, night, and late night periods of the day. To address these challenges, we propose a Dual Period-Varying Preference modeling (DPVP) for takeaway recommendation. Specifically, we design a dual interaction-aware module, aiming to capture users' dual preferences based on their interactions with stores and foods. Moreover, to model various preferences in different time periods of the day, we propose a time-based decomposition module as well as a time-aware gating mechanism. Extensive offline and online experiments demonstrate that our model outperforms state-of-the-art methods on real-world datasets and it is capable of modeling the dual period-varying preferences. Moreover, our model has been deployed online on Meituan Takeaway platform, leading to an average improvement in GMV (Gross Merchandise Value) of 0.70%.

Recently, a series of pioneer studies have shown the potency of pre-trained models in sequential recommendation, illuminating the path of building an omniscient unified pre-trained recommendation model for different downstream recommendation tasks. Despite these advancements, the vulnerabilities of classical recommender systems also exist in pre-trained recommendation in a new form, while the security of pre-trained recommendation model is still unexplored, which may threaten its widely practical applications. In this study, we propose a novel framework for backdoor attacking in pre-trained recommendation. We demonstrate the provider of the pre-trained model can easily insert a backdoor in pre-training, thereby increasing the exposure rates of target items to target user groups. Specifically, we design two novel and effective backdoor attacks: basic replacement and prompt-enhanced, under various recommendation pre-training usage scenarios. Experimental results on real-world datasets show that our proposed attack strategies significantly improve the exposure rates of target items to target users by hundreds of times in comparison to the clean model. The source codes are released in https://github.com/wyqing20/APRec.

Knowledge graph embedding (KGE) aims to project both entities and relations in a knowledge graph (KG) into low-dimensional vectors. Indeed, existing KGs suffer from the data imbalance issue, i.e., entities and relations conform to a long-tail distribution, only a small portion of entities and relations occur frequently, while the vast majority of entities and relations only have a few training samples. Existing KGE methods assign equal weights to each entity and relation during the training process. Under this setting, long-tail entities and relations are not fully trained during training, leading to unreliable representations. In this paper, we propose WeightE, which attends differentially to different entities and relations. Specifically, WeightE is able to endow lower weights to frequent entities and relations, and higher weights to infrequent ones. In such manner, WeightE is capable of increasing the weights of long-tail entities and relations, and learning better representations for them. In particular, WeightE tailors bilevel optimization for the KGE task, where the inner level aims to learn reliable entity and relation embeddings, and the outer level attempts to assign appropriate weights for each entity and relation. Moreover, it is worth noting that our technique of applying weights to different entities and relations is general and flexible, which can be applied to a number of existing KGE models. Finally, we extensively validate the superiority of WeightE against various state-of-the-art baselines.

The teacher-free online Knowledge Distillation (KD) aims to train an ensemble of multiple student models collaboratively and distill knowledge from each other. Although existing online KD methods achieve desirable performance, they often focus on class probabilities as the core knowledge type, ignoring the valuable feature representational information. We present a Mutual Contrastive Learning (MCL) framework for online KD. The core idea of MCL is to perform mutual interaction and transfer of contrastive distributions among a cohort of networks in an online manner. Our MCL can aggregate cross-network embedding information and maximize the lower bound to the mutual information between two networks. This enables each network to learn extra contrastive knowledge from others, leading to better feature representations, thus improving the performance of visual recognition tasks. Beyond the final layer, we extend MCL to intermediate layers and perform an adaptive layer-matching mechanism trained by meta-optimization. Experiments on image classification and transfer learning to visual recognition tasks show that layer-wise MCL can lead to consistent performance gains against state-of-the-art online KD approaches. The superiority demonstrates that layer-wise MCL can guide the network to generate better feature representations. Our code is publicly avaliable at https://github.com/winycg/L-MCL.

In recent years, graph neural networks (GNNs)-based methods have been widely adopted for heterogeneous graph (HG) embedding, due to their power in effectively encoding rich information from a HG into the low-dimensional node embeddings. However, previous works usually easily fail to fully leverage the inherent heterogeneity and rich semantics contained in the complex local structures of HGs. On the one hand, most of the existing methods either inadequately model the local structure under specific semantics, or neglect the heterogeneity when aggregating information from the local structure. On the other hand, representations from multiple semantics are not comprehensively integrated to obtain node embeddings with versatility. To address the problem, we propose aHeterogeneous Graph Neural Networkfor HG embeddingwithin a Multi-View representation learning framework(named MV-HetGNN), which consists of a view-specific ego graph encoder and auto multi-view fusion layer. MV-HetGNN thoroughly learns complex heterogeneity and semantics in the local structure to generate comprehensive and versatile node representations for HGs. Extensive experiments on three real-world HG datasets demonstrate the significant superiority of our proposed MV-HetGNN compared to the state-of-the-art baselines in various downstream tasks, e.g., node classification, node clustering, and link prediction.

Nowadays, Knowledge graphs (KGs) have been playing a pivotal role in AI-related applications. Despite the large sizes, existing KGs are far from complete and comprehensive. In order to continuously enrich KGs, automatic knowledge construction and update mechanisms are usually utilized, which inevitably bring in plenty of noise. However, most existing knowledge graph embedding (KGE) methods assume that all the triple facts in KGs are correct, and project both entities and relations into a low-dimensional space without considering noise and knowledge conflicts. This will lead to low-quality and unreliable representations of KGs. To this end, in this paper, we propose a general multi-task reinforcement learning framework, which can greatly alleviate the noisy data problem. In our framework, we exploit reinforcement learning for choosing high-quality knowledge triples while filtering out the noisy ones. Also, in order to take full advantage of the correlations among semantically similar relations, the triple selection processes of similar relations are trained in a collective way with multi-task learning. Moreover, we extend popular KGE models TransE, DistMult, ConvE and RotatE with the proposed framework. Finally, the experimental validation shows that our approach is able to enhance existing KGE models and can provide more robust representations of KGs in noisy scenarios.

Recently, recommender systems based on knowledge graphs (KGs) have become a popular research direction. Graph neural network (GNN) is the key technology of KG-based recommendation systems. However, existing GNNs have a significant flaw: They cannot explicitly model users’ intent in recommendations. Intent plays an essential role in users’ behaviors. For example, users may first generate an intent to purchase a certain group of items and then select a specific item from the group based on their preferences. Therefore, explicitly modeling intent has a positive significance for improving recommendation performance and providing explanations for recommendations. In this article, we propose a new model called Topic-aware Intention Network (TIN) for explainable recommendations with KGs. TIN models user representations from both preference and intent views. Specifically, we design a relational attention graph neural network to selectively aggregate information in KG to learn user preferences, and we propose a knowledge-enhanced topic model to learn user intent, which is viewed as topics hidden in user behavior sequences. Finally, we obtain the user representation by fusing user preference and intent through an attention network. The experimental results show that our proposed model outperforms the state-of-the-art methods and can generate reasonable explanations for the recommendation results.

Learning a disentangled representation is still a challenge in the field of the interpretability of generative adversarial networks (GANs). This paper proposes a generic method to modify a traditional GAN into an interpretable GAN, which ensures that filters in an intermediate layer of the generator encode disentangled localized visual concepts. Each filter in the layer is supposed to consistently generate image regions corresponding to the same visual concept when generating different images. The interpretable GAN learns to automatically discover meaningful visual concepts without any annotations of visual concepts. The interpretable GAN enables people to modify a specific visual concept on generated images by manipulating feature maps of the corresponding filters in the layer. Our method can be broadly applied to different types of GANs. Experiments have demonstrated the effectiveness of our method.

We proposed a Prior Gradient Mask Guided Pruning-aware Fine-Tuning (PGMPF) framework to accelerate deep Convolutional Neural Networks (CNNs). In detail, the proposed PGMPF selectively suppresses the gradient of those ”unimportant” parameters via a prior gradient mask generated by the pruning criterion during fine-tuning. PGMPF has three charming characteristics over previous works: (1) Pruning-aware network fine-tuning. A typical pruning pipeline consists of training, pruning and fine-tuning, which are relatively independent, while PGMPF utilizes a variant of the pruning mask as a prior gradient mask to guide fine-tuning, without complicated pruning criteria. (2) An excellent tradeoff between large model capacity during fine-tuning and stable convergence speed to obtain the final compact model. Previous works preserve more training information of pruned parameters during fine-tuning to pursue better performance, which would incur catastrophic non-convergence of the pruned model for relatively large pruning rates, while our PGMPF greatly stabilizes the fine-tuning phase by gradually constraining the learning rate of those ”unimportant” parameters. (3) Channel-wise random dropout of the prior gradient mask to impose some gradient noise to fine-tuning to further improve the robustness of final compact model. Experimental results on three image classification benchmarks CIFAR10/ 100 and ILSVRC-2012 demonstrate the effectiveness of our method for various CNN architectures, datasets and pruning rates. Notably, on ILSVRC-2012, PGMPF reduces 53.5% FLOPs on ResNet-50 with only 0.90% top-1 accuracy drop and 0.52% top-5 accuracy drop, which has advanced the state-of-the-art with negligible extra computational cost.

We present a collaborative learning method called Mutual Contrastive Learning (MCL) for general visual representation learning. The core idea of MCL is to perform mutual interaction and transfer of contrastive distributions among a cohort of networks. A crucial component of MCL is Interactive Contrastive Learning (ICL). Compared with vanilla contrastive learning, ICL can aggregate cross-network embedding information and maximize the lower bound to the mutual information between two networks. This enables each network to learn extra contrastive knowledge from others, leading to better feature representations for visual recognition tasks. We emphasize that the resulting MCL is conceptually simple yet empirically powerful. It is a generic framework that can be applied to both supervised and self-supervised representation learning. Experimental results on image classification and transfer learning to object detection show that MCL can lead to consistent performance gains, demonstrating that MCL can guide the network to generate better feature representations. Code is available at https://github.com/winycg/MCL.

Multivariate Time Series (MTS) forecasting plays a vital role in a wide range of applications. Recently, Spatial-Temporal Graph Neural Networks (STGNNs) have become increasingly popular MTS forecasting methods due to their state-of-the-art performance. However, recent works are becoming more sophisticated with limited performance improvements. This phenomenon motivates us to explore the critical factors of MTS forecasting and design a model that is as powerful as STGNNs, but more concise and efficient. In this paper, we identify the indistinguishability of samples in both spatial and temporal dimensions as a key bottleneck, and propose a simple yet effective baseline for MTS forecasting by attaching Spatial and Temporal IDentity information (STID), which achieves the best performance and efficiency simultaneously based on simple Multi-Layer Perceptrons (MLPs). These results suggest that we can design efficient and effective models as long as they solve the indistinguishability of samples, without being limited to STGNNs.

Recent years have witnessed remarkable progress on knowledge graph embedding (KGE) methods to learn the representations of entities and relations in static knowledge graphs (SKGs). However, knowledge changes over time. In order to represent the facts happening in a specific time, temporal knowledge graph (TKG) embedding approaches are put forward. While most existing models ignore the independence of semantic and temporal information. We empirically find that current models have difficulty distinguishing representations of the same entity or relation at different timestamps. In this regard, we propose a TimeLine-Traced Knowledge Graph Embedding method (TLT-KGE) for temporal knowledge graph completion. TLT-KGE aims to embed the entities and relations with timestamps as a complex vector or a quaternion vector. Specifically, TLT-KGE models semantic information and temporal information as different axes of complex number space or quaternion space. Meanwhile, two specific components carving the relationship between semantic and temporal information are devised to buoy the modeling. In this way, the proposed method can not only distinguish the independence of the semantic and temporal information, but also establish a connection between them. Experimental results on the link prediction task demonstrate that TLT-KGE achieves substantial improvements over state-of-the-art competitors. The source code will be available on https://github.com/zhangfw123/TLT-KGE.

Few-shot knowledge graph completion (FKGC) has become a new research focus in the field of knowledge graphs in recent years, which aims to predict the missing links for relations that only have a few associative triples. Existing models attempt to solve the problem via learning entity and relation representations. However, the limited training data severely hinders the performance of existing models. To this end, we propose to solve the FKGC problem with the data augmentation technique. Specifically, we perform data augmentation from two perspectives, i.e., inter-task view and intra-task view. The former generates new tasks for FKGC, while the latter enriches the support or query set for an individual task. It is worth noting that the proposed framework can be applied to a number of existing FKGC models. Experimental evaluation on two public datasets indicates our model is capable of achieving substantial improvements over baselines.

Current Knowledge Distillation (KD) methods for semantic segmentation often guide the student to mimic the teacher's structured information generated from individual data samples. However, they ignore the global semantic relations among pixels across various images that are valuable for KD. This paper proposes a novel Cross-Image Relational KD (CIRKD), which focuses on transferring structured pixel-to-pixel and pixel-to-region relations among the whole images. The motivation is that a good teacher network could construct a well-structured feature space in terms of global pixel dependencies. CIRKD makes the student mimic better structured semantic relations from the teacher, thus improving the segmentation performance. Experimental results over Cityscapes, CamVid and Pascal VOC datasets demonstrate the effectiveness of our proposed approach against state-of-the-art distillation methods. The code is available at https://github.com/winycg/CIRKD.

Existing works often focus on reducing the architecture redundancy for accelerating image classification but ignore the spatial redundancy of the input image. This paper proposes an efficient image classification pipeline to solve this problem. We first pinpoint task-aware regions over the input image by a lightweight patch proposal network called AnchorNet. We then feed these localized semantic patches with much smaller spatial redundancy into a general classification network. Unlike the popular design of deep CNN, we aim to carefully design the Receptive Field of AnchorNet without intermediate convolutional paddings. This ensures the exact mapping from a high-level spatial location to the specific input image patch. The contribution of each patch is interpretable. Moreover, AnchorNet is compatible with any downstream architecture. Experimental results on ImageNet show that our method outperforms SOTA dynamic inference methods with fewer inference costs. Our code is available at https://github.com/winycg/AnchorNet.

This paper addresses the problem of scheduling real-time wireless flows under dynamic network conditions and general traffic patterns. The objective is to maximize the fraction of packets of each flow to be delivered within their deadlines, referred to as timely-throughput. The scheduling problem under restrictive frame-based traffic models or greedy maximal scheduling schemes like LDF has been extensively studied so far, but scheduling algorithms to provide deadline guarantees on packet delivery for general traffic under dynamic network conditions are very limited. We propose two scheduling algorithms using deep reinforcement learning approach to optimize timely-throughput for general traffic in dynamic wireless networks: RL-Centralized scheduling algorithm and RL-Decentralized scheduling algo-rithm. Specifically, we formulate the centralized scheduling problem as a Markov Decision Process (MDP) and a multi-environments double deep Q-network (ME-DDQN) structure is proposed to adapt to the dynamic network conditions. The decentralized scheduling problem is formulated as a Partially Observable Markov Decision Process (POMDP) and an expert-apprentice centralized training and decentralized execution (EA-CTDE) structure is designed to accelerate the training speed and achieve the optimal timely-throughput. The extensive results show that the proposed scheduling algorithms converge fast and adapt well to network dynamics with superior performance compared to baseline policies. Finally, experimental tests confirm simulation results and also show that the proposed algorithms are feasible in practice on resource limited platforms.

Multivariate Time Series (MTS) forecasting plays a vital role in a wide range of applications. Recently, Spatial-Temporal Graph Neural Networks (STGNNs) have become increasingly popular MTS forecasting methods. STGNNs jointly model the spatial and temporal patterns of MTS through graph neural networks and sequential models, significantly improving the prediction accuracy. But limited by model complexity, most STGNNs only consider short-term historical MTS data, such as data over the past one hour. However, the patterns of time series and the dependencies between them (i.e., the temporal and spatial patterns) need to be analyzed based on long-term historical MTS data. To address this issue, we propose a novel framework, in which STGNN is Enhanced by a scalable time series Pre-training model (STEP). Specifically, we design a pre-training model to efficiently learn temporal patterns from very long-term history time series (e.g., the past two weeks) and generate segment-level representations. These representations provide contextual information for short-term time series input to STGNNs and facilitate modeling dependencies between time series. Experiments on three public real-world datasets demonstrate that our framework is capable of significantly enhancing downstream STGNNs, and our pre-training model aptly captures temporal patterns.

Predicting the future trajectories of multiple agents is essential for various applications in real life, such as surveillance systems, autonomous driving and social robots. The trajectory prediction task is influenced by many factors, including the individual historical trajectory, interactions between agents and fuzzy nature of an agent's motion. While existing methods have made great progress on the topic of trajectory prediction, they treat all the information uniformly, which limits the sufficiency of using information. To this end, in this paper, we propose to regard all the information in a two-level hierarchical view. Particularly, the first-level view is the inter-trajectory view. In this level, we observe that the difficulty to predict different trajectory samples is different. We define trajectory difficulty and train the proposed model in an "easy-to-hard'' schema. The second-level view is the intra-trajectory level. We find the influencing factors for a particular trajectory can be divided into two parts. The first part is global features, which keep stable within a trajectory, i.e., the expected destination. The second part is local features, which change over time, i.e., the current position. We believe that the two types of information should be handled in different ways. The hierarchical view is beneficial to take full advantage of the information in a fine-grained way. Experimental results validate the effectiveness of the proposed model.

We all depend on mobility, and vehicular transportation affects the daily lives of most of us. Thus, the ability to forecast the state of traffic in a road network is an important functionality and a challenging task. Traffic data is often obtained from sensors deployed in a road network. Recent proposals on spatial-temporal graph neural networks have achieved great progress at modeling complex spatial-temporal correlations in traffic data, by modeling traffic data as a diffusion process. However, intuitively, traffic data encompasses two different kinds of hidden time series signals, namely the diffusion signals and inherent signals. Unfortunately, nearly all previous works coarsely consider traffic signals entirely as the outcome of the diffusion, while neglecting the inherent signals, which impacts model performance negatively. To improve modeling performance, we propose a novel Decoupled Spatial-Temporal Framework (DSTF) that separates the diffusion and inherent traffic information in a data-driven manner, which encompasses a unique estimation gate and a residual decomposition mechanism. The separated signals can be handled subsequently by the diffusion and inherent modules separately. Further, we propose an instantiation of DSTF, Decoupled Dynamic Spatial-Temporal Graph Neural Network (D 2 STGNN), that captures spatial-temporal correlations and also features a dynamic graph learning module that targets the learning of the dynamic characteristics of traffic networks. Extensive experiments with four real-world traffic datasets demonstrate that the framework is capable of advancing the state-of-the-art.

In this study, a new learning control scheme is designed to investigate the stability of a multilayer supply chain network (SCN) and to further improve the convergence speed of the nodes’ states of such a multilayer SCN. Specifically, a multilayer SCN model with three layers is first established and some practical constraints on the states of the proposed SCN model are involved and discussed. By taking the quantities of goods transmitted between different nodes as control inputs, a new kind of learning control scheme is subsequently proposed to discuss the stability of the nodes’ states within the SCN. It is further shown that the convergence speed of nodes’ states with this scheme is faster than that yielded by using some traditional schemes. The contributions of our scheme are twofold: 1) it can save the limited control resource and 2) it can improve the convergence speeds of the states of all nodes. Numerical simulations are finally given to illustrate the effectiveness and advantages of the designed learning scheme.

Previous Online Knowledge Distillation (OKD) often carries out mutually exchanging probability distributions, but neglects the useful representational knowledge. We there-fore propose Multi-view Contrastive Learning (MCL) for OKD to implicitly capture correlations of feature embeddings encoded by multiple peer networks, which provide various views for understanding the input data instances. Benefiting from MCL, we can learn a more discriminative representation space for classification than previous OKD methods. Experimental results on image classification demonstrate that our MCL-OKD outperforms other state-of-the-art OKD methods by large margins without sacrificing additional inference cost. Codes are available at https://github.com/winycg/MCL-OKD.

This paper considers a time-critical task allocation problem in a distributed multi-UAV system. Existing distributed task allocation algorithms trend to increase communication overhead due to the resolution of numerous task-bundle conflicts between UAVs and easily trap into local optimum with greedy strategy. In this work, we propose a novel distributed task allocation method. First, tasks are divided into multiple clusters, and then UAVs build their task bundles from separate task clusters to avoid conflicts between them, thereby reducing communication overhead. Second, to increase the exploratory ability, an improved ant colony optimization algorithm is proposed to achieve task allocation of UAVs from their corresponding task clusters, instead of using greedy-based strategy. Moreover, an inter-cluster adjustment mechanism is proposed to solve unassigned tasks in task clusters to improve task assignment ratio with low communication overhead, which has been verified in our simulations. Extensive simulation results confirm that our method can achieve efficient task allocation solution with high task assignment ratio and low communication overhead when compared with the state-of-the-art algorithms.

Knowledge distillation often involves how to define and transfer knowledge from teacher to student effectively. Although recent self-supervised contrastive knowledge achieves the best performance, forcing the network to learn such knowledge may damage the representation learning of the original class recognition task. We therefore adopt an alternative self-supervised augmented task to guide the network to learn the joint distribution of the original recognition task and self-supervised auxiliary task. It is demonstrated as a richer knowledge to improve the representation power without losing the normal classification capability. Moreover, it is incomplete that previous methods only transfer the probabilistic knowledge between the final layers. We propose to append several auxiliary classifiers to hierarchical intermediate feature maps to generate diverse self-supervised knowledge and perform the one-to-one transfer to teach the student network thoroughly. Our method significantly surpasses the previous SOTA SSKD with an average improvement of 2.56% on CIFAR-100 and an improvement of 0.77% on ImageNet across widely used network pairs. Codes are available at https://github.com/winycg/HSAKD.

Filter pruning is widely used to reduce the computation of deep learning, enabling the deployment of Deep Neural Networks (DNNs) in resource-limited devices. Conventional Hard Filter Pruning (HFP) method zeroizes pruned filters and stops updating them, thus reducing the search space of the model. On the contrary, Soft Filter Pruning (SFP) simply zeroizes pruned filters, keeping updating them in the following training epochs, thus maintaining the capacity of the network. However, SFP, together with its variants, converges much slower than HFP due to its larger search space. Firstly, we generalize SFP-based methods and HFP to analyze their characteristics. Then we propose a Gradually Hard Filter Pruning (GHFP) method to smoothly switch from SFP-based methods to HFP during training and pruning, thus maintaining a large search space at first, gradually reducing the capacity of the model to ensure a moderate convergence speed. Furthermore, we view filter pruning as dimension reduction and propose a novel dimension reduction block integrated into GHFP to significantly outperform other methods by a moderate margin.

Measuring the similarity between trajectories is fundamental to many location-aware applications. However, the trajectory data collected in the real world suffer from the low-quality problem caused by non-uniform sampling rates and noises, which significantly affects the accuracy of similarity measurement. Traditional pairwise point-matching methods are susceptible to non-uniform sampling rates inherently, since they assume the consistent sampling rate. Although the recurrent neural network (RNN) based methods have addressed this problem by complementing trajectory data, they have the drawback of predicting positions conditioned on the historical data generated by the model itself, which could lead to accumulated bias during the inference stage. In this paper, we propose a novel generative model to address the important issue of low-quality trajectory data based on single image super-resolution. The trajectory similarity is thus computed using trajectory images instead of the trajectory sequential data. By utilizing the images to represent trajectories, we effectively overcome the issues encountered by existing methods. Extensive experimental results using real-world datasets demonstrate that our method outperforms existing methods in terms of accuracy.

Channel pruning has demonstrated its effectiveness in compressing ConvNets. In many related arts, the importance of an output feature map is only determined by its associated filter. However, these methods ignore a small part of weights in the next layer which disappears as the feature map is removed. They ignore the phenomenon of weight dependency. Besides, many pruning methods use only one criterion for evaluation and find a sweet spot of pruning structure and accuracy in a trial-and-error fashion, which can be time-consuming. In this paper, we proposed a channel pruning algorithm via multi-criteria based on weight dependency, CPMC, which can compress a pre-trained model directly. CPMC defines channel importance in three aspects, including its associated weight value, computational cost, and parameter quantity. According to the phenomenon of weight dependency, CPMC gets channel importance by assessing its associated filter and the corresponding partial weights in the next layer. Then CPMC uses global normalization to achieve cross-layer comparison. Finally, CPMC removes less important channels by global ranking. CPMC can compress various CNN models, including VGGNet, ResNet, and DenseNet on various image classification datasets. Extensive experiments have shown CPMC outperforms the others significantly.

In this paper, we address the problem of building adaptive network coding coefficients under dynamic network conditions (e.g., varying link quality and changing number of relays). In existing linear network coding solutions including deterministic network coding and random linear network coding, coding coefficients are set by a heuristic or randomly chosen from a Galois field with equal probability, which can not adapt to dynamic network conditions with good decoding performance. We propose INCdeep, an adaptive Intelligent Network Coding with Deep Reinforcement Learning. Specifically, we formulate a coding coefficients selection problem where network variations can be automatically and continuously expressed as the state transitions of a Markov decision process (MDP). The key advantage is that INCdeep is able to learn and dynamically adjust the coding coefficients for the source node and each relay node according to ongoing network conditions, instead of randomly. The results show that INCdeep has generalization ability that adapts well in dynamic scenarios where link quality is changing fast, and it converges fast in the training process. Compared with the benchmark coding algorithms, INCdeep shows superior performance, including higher decoding probability and lower coding overhead through simulations and experiments.

This paper addresses the problem of UAV positioning and resource allocation under dynamic network conditions and under instantaneous communication demands in FANETs. We propose MPRdeep, an adaptive, deep reinforcement learning (DRL) approach considering several QoS requirements concurrently. MPRdeep learns to optimize relay UAVs’ positions and forwarding probabilities to minimize reliability-achieving delay and reliability-achieving energy consumption. The key advantage is that MPRdeep is able to learn and dynamically adjust the node positioning and resource allocation according to ongoing network conditions. The results show that MPRdeep converges fast and has generalization ability that adapts well under dynamic network conditions and dynamic locations of users. Compared with baseline methods, MPRdeep shows superior performance in terms of lower reliability-achieving delay and lower reliability-achieving energy consumption via simulations and experiments.

We propose a simple yet effective method to reduce the redundancy of DenseNet by substantially decreasing the number of stacked modules by replacing the original bottleneck by our SMG module, which is augmented by local residual. Furthermore, SMG module is equipped with an efficient two-stage pipeline, which aims to DenseNet-like architectures that need to integrate all previous outputs, i.e., squeezing the incoming informative but redundant features gradually by hierarchical convolutions as a hourglass shape and then exciting it by multi-kernel depthwise convolutions, the output of which would be compact and hold more informative multi-scale features. We further develop a forget and an update gate by introducing the popular attention modules to implement the effective fusion instead of a simple addition between reused and new features. Due to the Hybrid Connectivity (nested combination of global dense and local residual) and Gated mechanisms, we called our network as the HCGNet. Experimental results on CIFAR and ImageNet datasets show that HCGNet is more prominently efficient than DenseNet, and can also significantly outperform state-of-the-art networks with less complexity. Moreover, HCGNet also shows the remarkable interpretability and robustness by network dissection and adversarial defense, respectively. On MS-COCO, HCGNet can consistently learn better features than popular backbones.

Convolutional neural networks (ConvNets) are widely used in real life. People usually use ConvNets which pre-trained on a fixed number of classes. However, for different application scenarios, we usually do not need all of the classes, which means ConvNets are redundant when dealing with these tasks. This paper focuses on the redundancy of ConvNet channels. We proposed a novel idea: using an interpretable manner to find the most important channels for every single class (dissect), and dynamically run channels according to classes in need (reconstruct). For VGG16 pre-trained on CIFAR-10, we only run 11\% parameters for two-classes sub-tasks on average with negligible accuracy loss. For VGG16 pre-trained on ImageNet, our method averagely gains 14.29\% accuracy promotion for two-classes sub-tasks. In addition, analysis show that our method captures some semantic meanings of channels, and uses the context information more targeted for sub-tasks of ConvNets.

Hypothesis-oriented Multiple Hypothesis Tracking (HOMHT) recursively generates hypotheses on the origins of measurements and manages them, therefore it is computationally intensive. To speed up HOMHT for tracking hundreds of targets in real time, we propose a parallel implementation of this algorithm which distributes hypotheses into independent worker threads residing in multiple CPU cores. The implementation in this paper is based on object-oriented programming: each hypothesis object manages its target data all by itself and the generation and pruning of a hypothesis is achieved by its copy constructor and destructor functions. We evaluate this method by tracking 150 targets through 3 heterogeneous sensors in real-time with 32-best hypotheses running in 1, 2, 4, 8, 16 and 32 worker threads respectively. The results validate the method's scalability in which measurement fusion latency is approximately inversely proportional to worker thread count. We also make a pressure test by tracking 500 targets through 3 sensors, and HOMHT is able to run concurrently in real-time with 32 worker threads.

Existing approaches to improve the performances of convolutional neural networks by optimizing the local architectures or deepening the networks tend to increase the size of models significantly. In order to deploy and apply the neural networks to edge devices which are in great demand, reducing the scale of networks is quite crucial. However, It is easy to degrade the performance of image processing by compressing the networks. In this paper, we propose a method which is suitable for edge devices while improving the efficiency and effectiveness of inference. The joint decision of multiparticipants, mainly contain multi-layers and multi-networks, can achieve higher classification accuracy (0.26% on CFAR-10 and 4.49% on CFAR-100 at most) with similar total number of parameters for classical convolutional neural networks.

Network pruning is widely used to compress Deep Neural Networks (DNNs). The Soft Filter Pruning (SFP) method zeroizes the pruned filters during training while updating them in the next training epoch. Thus the trained information of the pruned filters is completely dropped. To utilize the trained pruned filters, we proposed a SofteR Filter Pruning (SRFP) method and its variant, Asymptotic SofteR Filter Pruning (ASRFP), simply decaying the pruned weights with a monotonic decreasing parameter. Our methods perform well across various networks, datasets and pruning rates, also transferable to weight pruning. On ILSVRC-2012, ASRFP prunes 40% of the parameters on ResNet-34 with 1.63% top-1 and 0.68% top-5 accuracy improvement. In theory, SRFP and ASRFP are an incremental regularization of the pruned filters. Besides, We note that SRFP and ASRFP pursue better results while slowing down the speed of convergence.

The prevalent adoptions of GPS-enabled devices have witnessed an explosion of various location-based services which produces a huge amount of trajectories monitoring the individuals' movements. In this paper, we tackle Trajectory-User Link (TUL) problem, which identifies humans' movement patterns and links trajectories to the users who generated them. Existing solutions on TUL problem employ recurrent neural networks and variational autoencoder methods, which face the bottlenecks in the case of excessively long trajectories and fragmentary users' movements. However, these are common characteristics of trajectory data in reality, leading to performance degradation of the existing models. In this paper, we propose an end-to-end attention recurrent neural learning framework, called TULAR (Trajectory-User Link with Attention Recurrent Networks), which focus on selected parts of the source trajectories when linking. TULAR introduce the Trajectory Semantic Vector (TSV) via unsupervised location representation learning and recurrent neural networks, by which to reckon the weight of parts of source trajectory. Further, we employ three attention scores for the weight measurements. Experiments are conducted on two real world datasets and compared with several existing methods, and the results show that TULAR yields a new state-of-the-art performance. Source code is public available at GitHub: https://github.com/taos123/TULAR.

In this paper we propose to model high and low frequencies of segmentation map, based on the observation that the map can be seen as a mixture of different frequencies. Based on the sparsity of high frequencies and local similarity of low frequencies, we design special building blocks and further a novel High and Low frequency Network (HLNet) with two branches based on FCN to predict high and low frequencies of the segmentation map, respectively. Specifically, we design a high frequency branch with a small kernel size and high-resolution features to predict a sparse high frequency component. Mean-while, a low frequency branch with similarity computing and low-resolution features is employed to predict a low frequency component. On top of two branches, we combine two different frequency components to generate final result for scene parsing. We empirically demonstrate that the designed model achieves superior performance 44.07% on ADE20K, and 80.14% mIoU on Cityscapes datasets.

Trajectory-user linking (TUL), whereby a trajectory is linked to its owner in location-based social networks, is a fundamental and critical task in spatio-temporal data mining. It plays a key role in personalized recommendation, anomaly detection, and semantic trajectory mining. Existing methods for TUL are either rule-based methods, which link trajectories and users based on conventional trajectory similarities, or learning-based methods, which learn a classification model to map trajectories to their owners. However, rule-based methods ignore the semantic information in the trajectory sequence, and learning-based methods require retraining the model each time a new user is added. In this paper, we propose a Siamese network-based model for trajectory-user linking (TULSN), which uses a Siamese network to capture semantic information in the trajectory, and instead of retraining the model, it requires only a few labeled trajectories per user to identify the user category of the trajectory. The experimental results show that the TULSN outperforms existing baselines and state-of-the-art methods on real-world datasets.

Latest algorithms for automatic neural architecture search perform remarkably but few of them can effectively design the number of channels for convolutional neural networks and consume less computational efforts. In this paper, we propose a method for efficient automatic search which is special to the widths of networks instead of the connections within neural architectures. Our method, functionally incremental search based on function-preserving, will explore the number of channels for almost any convolutional neural network rapidly while controlling the number of parameters and even the amount of computations (FLOPs). On CIFAR-10 and CIFAR-100 classification, our method using minimal computational resources (0.41 ~ 1.29 GPU-days) can discover more effective rules of the widths of networks to improve the accuracy (a ~ 1.08 on CIFAR-10 and b ~ 2.33 on CIFAR-100) with fewer number of parameters.

The finite-difference dynamical core based on the equal-interval latitude-longitude mesh has been widely used for numerical simulations of the Atmospheric General Circulation Model (AGCM). Previous work utilizes different filtering schemes to alleviate the instability problem incurred by the unequal physical spacing at different latitudes, but they all incur high communication and computation overhead and become a scaling bottleneck. This paper proposes a new leap-format finite-difference computing scheme. It generalizes the usual finite-difference format with adaptive wider intervals and is able to maintain the computational stability in the grid updating. Therefore, the costly filtering scheme is eliminated. The new scheme is parallelized with a shifting communication method and implemented with fine communication optimizations based on a 3D decomposition. With the proposed leap-format computation scheme, the communication overhead of the AGCM is significantly reduced and good load balance is exhibited. The simulation results verify the correctness of the new leap-format scheme. The new scheme achieves the speed of 16.6 simulation-year-per-day (SYPD) and up to 3.3x speedup over the latest implementation.

The mobile robotic network consisting multiple robotic devices such as unmanned aerial vehicles (UAVs) is a high-speed mobile wireless network. Existing mobile ad hoc protocols cannot meet the demands of mobile robotic networks due to intermittently connected links and frequent topology changes. This paper proposes a deep reinforcement learning based adaptive and reliable routing protocol, ARdeep. We formulate routing decisions with a Markov Decision Process model to automatically characterize the network variations. To better infer network environment, the link status is considered when making routing decisions. Simulation results demonstrate that ARdeep outperforms the existing good performing QGeo and conventional GPSR.

In many industrial wireless networks, sensor nodes sense an event and then generate related data packets to transmit. In such event-driven networks, contention medium access control (MAC) protocols such as 802.11 DCF are widely used to address the burst traffic. However, due to the fixed probability distribution of contention time-slot selection (PDCS), traditional methods may cause severe collisions when a great number of packets are being transmitted in a short period. To deal with this problem, a MAC protocol that dynamically adjusts PDCS and contention window simultaneously is proposed in this paper. This method exploits the spatial and temporal relationship of packet generation among source nodes in industrial wireless networks. In this method, sensor nodes related to the same event source converge to the optimal PDCS quickly and collisions are significantly reduced. Delay performances of this method are evaluated through simulations. According to the simulation, our protocol outperforms current MAC protocols in access delay by about 20% in event-driven networks.

A network with reliable and rapid communication is critical for Unmanned Aerial Vehicles (UAVs). Flying Ad Hoc Networks (FANETs) consisting of UAVs is a new paradigm of wireless communication . However, the highly dynamic topology of FANETs and limited energy of UAVs have brought great challenges to the routing design of FANETs. It is difficult for existing routing protocols for Mobile Ad Hoc Networks (MANETs) and Vehicular Ad Hoc Networks (VANETs) to adapt the high dynamics of FANETs. Moreover, few of existing routing protocols simultaneously meet the requirement of low delay and low energy consumption of FANETs. This paper proposes a novel Q-learning based Multi-objective optimization Routing protocol for FANETs to provide low-delay and low-energy service guarantees. Most of existing Q-learning based protocols use a fixed value for the Q-learning parameters. In contrast, Q-learning parameters can be adaptively adjusted in the proposed protocol to adapt to the high dynamics of FANETs. In addition, a new exploration and exploitation mechanism is also proposed to explore some undiscovered potential optimal routing path while exploiting the acquired knowledge. Instead of using past neighbor relationships, the proposed method re-estimates neighbor relationships in the routing decision process to select the more reliable next hop. Simulation results show that the proposed method can provide higher packet arrival ratio, lower delay and energy consumption than existing good performing Q-learning based routing method.

Communications in industrial applications like wireless factory automation demands different timing requirements. Providing timely medium access of the critical traffic and its prioritization over regular traffic is a significant challenge in industrial wireless networks. Successive Interference Cancellation (SIC) technique is an effective way to decrease access delay by allowing multiple transmissions concurrently. A series of novel Medium Access Control (MAC) protocols are proposed to differentiate access delay for various traffic types or only exploit SIC for unique traffic type. However, to the best of our knowledge, this work is the first priority-based distributed MAC protocol that employs SIC (PDMAC-SIC) to provide low delay and accommodate different types of traffic for industrial wireless networks. There are two major contributions of our work: first, an extra power contention procedure other than traditional RTS/CTS contention in CSMA/CA is introduced in our PDMAC-SIC. This power contention procedure allows multiple transmitters to access the same channel simultaneously and thus access delay is decreased. Second, PDMAC-SIC is modeled by Markov chain and then the access delay is minimized by optimizing the size of power contention window. Our analytical model is verified through simulation. Results reveal that PDMAC-SIC performs better on access delay and packet loss rate than the existing good performing priority based CSMA/CA.

Nowadays in many application scenarios of Internet of Things (IoT), low latency is achieved at the cost of computing-complexity which is beyond the capabilities of IoT devices. Offloading the computing intensive tasks to more powerful edge devices is expected to provide new generation computing-intensive and delay-sensitive services. In the three hierarchy architecture user/IoT-edge-cloud, private and secure mutual authentication are necessary between user, IoT device, and edge device. However, in the emerging computing paradigms, such as mobile transparent computing, edge computing, fog computing, and several threats, such as edge device compromise, privacy leaking, and denial of service (DoS) might crash the security of the system. Here, we propose a lightweight anonymous mutual authentication scheme for ${n}$ -times computing offloading (CO) in IoT. In our novel scheme, through a smartcard as token and an edge device as a security proxy, a user is able to subscribe or renew ${n}$ -times CO service and consume it securely in daily use. Moreover, both IoT and edge devices authenticate each other anonymously without leaking user's sensitive information, which will preserve the privacy even when an edge device is comprised. Finally, our scheme is based on lightweight one-way hash function and MAC function, therefore the adversary is not able to perform a DoS attack. To evaluate the solution, a security analysis and a performance analysis are presented. Compared with similar schemes, our approach achieves all designed security features and achieves a $1.66\boldsymbol {\times }$ and $2.87\boldsymbol {\times }$ of computing speed on IoT and edge devices, respectably.

Wireless networks have become a very attractive solution for soft real-time data transport in the industry. For such technologies to carry real-time traffic, reliable bounds on end-to-end communication delays have to be ascertained to warrant a proper system behavior. As for legacy wired embedded and real-time networks, two main wireless multiple access methods can be leveraged: one is time division multiple access (TDMA), which follows a time-triggered paradigm, and the other is carrier sense multiple access with collision avoidance (CSMA/CA), which follows an event-triggered paradigm. This paper proposes an analytical comparison of the time behavior of two representative TDMA and CSMA/CA protocols in terms of the worst-case end-to-end delay. This worst-case delay is expressed in a probabilistic manner because our analytical framework captures the versatility of the wireless medium. Analytical delay bounds are obtained from delay distributions, which are compared to fine-grained simulation results. Exhibited study cases show that TDMA can offer smaller or larger worst-case bounds than CSMA/CA depending on its settings.

With the rapid development of marine technology, underwater sensor networks (UWSNs) are gradually evolving from research to practice in recent years. Practicability and reliability are two major concerns for routing protocols in UWSNs. As localization is not necessary in depth-based routing protocol (DBR), it has an outstanding practicability than other geographic routing protocols. However, the reliability is not well ensured. In this paper, we propose an innovative depth-based routing with network coding improving routing reliability while preserving the intrinsic distributed manner of DBR and introducing little time delay and energy cost. Moreover, a simple analytical performance model where ideal MAC is assumed is proposed to derive the analytical delivery ratio for our DBR-NC and DBR protocols. This analytical model is validated by simulation results. The extensive simulation results show that the proposed DBR-NC protocol outperforms (over 15%) the state of art DBR protocols in terms of packet delivery ratio. We also show that our DBR-NC will not introduce much extra delay and energy consumptions.

Wireless networks have become a very attractive solution for soft real-time data transport in the industry. For such technologies to carry real-time traffic, reliable bounds on end-to-end communication delays have to be ascertained to warrant a proper system behavior. As for legacy wired embedded and real-time networks, two main wireless multiple access methods can be leveraged: (i) time division multiple access (TDMA), which follows a time-triggered paradigm and (ii) Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA), which follows an event-triggered paradigm. This paper proposes an analytical comparison of the time behavior of two representative TDMA and CSMA/CA protocols in terms of worst-case end-to-end delay. This worst-case delay is expressed in a probabilistic manner because our analytical framework captures the versatility of the wireless medium. Analytical delay bounds are obtained from delay distributions, which are compared to fine-grained simulation results. Exhibited study cases show that TDMA can offer smaller or larger worst-case bounds than CSMA/CA depending on its settings.

Human tracking is one of the most important requirements for service mobile robots. Cameras and Laser Ranger Finders (LRFs) are usually used together for human tracking. But these kinds of solutions are too computationally expensive for most embedded processors on these robots as complex computer vision algorithms are needed to process large number of pixels. In this paper, we describe a method combining kinematic measurements from LRF mounted on the robot and Inertial Measurement Unit (IMU) carried by the target. These two types of sensors can calculate human's velocity and position independently, which are used as information for both indentifying and tracking the target. As pixels observed by LRF and IMU are 1D rather than 2D, our method requires much less computation and memory resources and can be implemented with low-performance embedded processors.

No abstract available.

Some previous work has shown the feasibility of Pedestrian Dead Reckoning (PDR) using a mobile phone, but the estimation of length walked is still a big challenge. In this paper, we propose a formula for estimating walking velocity, which can be integrated to get distance.

Abstract Link quality is one of the most important factors that affect the performance of wireless networks. In a densely deployed wireless network, continuous link quality monitoring consumes significant amount of energy and bandwidth at each node. In this paper, we propose a sensitivity model and a spatial correlation model that can be used to derive a set of deputy links to monitor, instead of monitoring all of the links in the network. The proposed scheme can improve energy efficiency of the link quality monitoring process. A greedy algorithm is presented to derive the deputy links set based on three different optimization objective functions. Performance of the proposed method is studied extensively and it is shown that the proposed method can save almost 90% energy in typical simulation scenarios than the method of monitoring all links. We also demonstrate that the energy consumption of the greedy deputy set‐based method is upper‐bounded. Copyright © 2010 John Wiley & Sons, Ltd.

Wireless sensor networks need energy-efficient mechanisms of performance measurement for various aspects of network design, optimization and management. In this paper, we take into account the unique data aggregation communication paradigm of wireless sensor networks: the network performance can be measured by exploiting whether application data from various sensor nodes reach the sink, without incurring any additional overhead of active probes or performance reports. Then we present a novel algorithm, which can infer sensor network topology and link loss performance simultaneously. Finally, we validate the algorithm through simulations and it exhibit ts good performance and scalability.

Power consumption has become a primary constraint of integrated circuit design. Many models have been proposed to evaluate dynamic and leakage power in every design level. However, how to accurately predict TPower, the total power with dynamic and leakage power included, for large-scale designs within reasonable time remains unsolved. In this paper, a new topic of vector extraction for power estimation is brought forward based on the distribution analysis of power consumption of different types. After extracted, a large number of vectors are compacted into much fewer without significant influence on the specified power property, which makes the application of accurate and fast simulator possible. For the purpose of validation, we use the method on average TPower vector extraction and obtain good experimental results.

Crosstalk issues in deep sub-micron (DSM) cause severe design validation and test problems. This paper addresses the problems of delay test considering crosstalkinduced effects, and proposes a non-robust delay test generation method based on single precise crosstalkinduced path delay fault (S-PCPDF) model. With some necessary static timing analysis, the target crosstalkinduced delay fault set was reduced. And the delay test generation for crosstalk-induced delay faults can be implemented by few alterations of non-robust path delay test generation algorithms whereas the timing information is only considered during the selection of target faults. Experimental results on ISCAS’89 and ITC’99 benchmark circuits showed that the proposed method does efficiently for circuits of reasonable sizes, and the CPU time is acceptable.

This paper presents a test resource partitioning technique based on an efficient single-output response compaction design called quotient compactor (q-Compactor). Some design theorems of quotient compactor are presented to achieve full diagnostics ability, minimize error cancellation and handle the X bits in the outputs of the CUT The quotient compactor can also be moved to the load-board to reduce the number of ATE channels required. Our experimental results on the ISCA S89 benchmark circuits and an MPEG 2 decoder SOC show that the proposed compaction scheme is very efficient.

Dual threshold voltage design is the most effective technique for reducing leakage current of integrated circuits. In this paper, we put forward an average leakage current macromodeling for dual-threshold circuits and propose two methods to conquer it, table-lookup based simulation and statistical analysis. The simulation is an efficient general-purposed method for leakage current estimation. For stochastic input circuits, a new statistical method is presented, which is much faster than the former one and can be used for very large circuits, because the method is based on input Signal probability and only one pass of circuit is needed. To validate these two methods, accurate leakage currents are also calculated using popular simulator HSPICE for comparison. Experiments on ISCAS85 and ISCAS89 benchmarks demonstrate that the errors between them can be limited under 4% and thousand folds of time to results can be saved.