Large vision-language models (LVLMs) have shown significant progress in video understanding, but they struggle to scale to long videos due to limited context windows. Existing methods reduce input dimensionality via frame sampling and feature compression, yet discard details and incur high computational cost for post-training. In contrast, retrieval-augmented generation (RAG) that indexes long videos for query retrieval and memory-based methods that maintain evolving long-term stores, offer a lighter and deployment-friendly solution. Nevertheless, they rely on shallow retrieval that selects only top-ranked segments and fails to integrate information across multiple relevant video episodes. Inspired by Multiple-Trace Theory in cognitive psychology, we revisit long video understanding from a probe-echo perspective, in which human episodic memories are activated and integrated in parallel. Building on this insight, we propose ProEchoMem, a cognitive-inspired framework that simulates the probe-echo mechanism: (1) Incremental Episodic Memory Construction builds structured knowledge graphs from video streams; (2) Probe-Driven Memory Activation generates probe signals from user queries to activate all stored traces simultaneously; (3) Memory Echo Synthesis integrates activated traces into a coherent and structured memory echo. Experiments on LongerVideos, LVBench, and cross-domain settings demonstrate the effectiveness of ProEchoMem, with multi-trace probing achieving up to 14.2% higher relevance and ablation studies validating the contribution of each module. The code is available at https://github.com/Applied-Machine-Learning-Lab/SIGIR26_ProEchoMem

Transformer structures have been widely used in sequential recommender systems (SRS). However, as user interaction histories increase, computational time and memory requirements also grow. This is mainly caused by the standard attention mechanism. Although there exist many methods employing efficient attention and SSM-based models, these approaches struggle to effectively model long sequences and may exhibit unstable performance on short sequences. To address these challenges, we design a sparse attention mechanism, BlossomRec, which models both long-term and short-term user interests through attention computation to achieve stable performance across sequences of varying lengths. Specifically, we categorize user interests in recommendation systems into long-term and short-term interests, and compute them using two distinct sparse attention patterns, with the results combined through a learnable gated output. Theoretically, it significantly reduces the number of interactions participating in attention computation. Extensive experiments on four public datasets demonstrate that BlossomRec, when integrated with state-of-the-art Transformer-based models, achieves comparable or even superior performance while significantly reducing memory usage, providing strong evidence of BlossomRec's efficiency and effectiveness. The code is available at https://github.com/Applied-Machine-Learning-Lab/WWW2026_BlossomRec.

Recent advances in generative artificial intelligence, particularly large language models (LLMs), have opened new opportunities for enhancing recommender systems (RecSys). Most existing LLM-based RecSys approaches operate in a discrete space, using vector-quantized tokenizers to align with the inherent discrete nature of language models. However, these quantization methods often result in lossy tokenization and suboptimal learning, primarily due to inaccurate gradient propagation caused by the non-differentiable argmin operation in standard vector quantization. Inspired by the emerging trend of embracing continuous tokens in language models, we propose ContRec, a novel framework that seamlessly integrates continuous tokens into LLM-based RecSys. Specifically, ContRec consists of two key modules: a σ-VAE Tokenizer, which encodes users/items with continuous tokens; and a Dispersive Diffusion module, which captures implicit user preference. The tokenizer is trained with a continuous Variational Auto-Encoder (VAE) objective, where three effective techniques are adopted to avoid representation collapse. By conditioning on the previously generated tokens of the LLM backbone during user modeling, the Dispersive Diffusion module performs a conditional diffusion process with a novel Dispersive Loss, enabling high-quality user preference generation through next-token diffusion. Finally, ContRec leverages both the textual reasoning output from the LLM and the latent representations produced by the diffusion model for Top-K item retrieval, thereby delivering comprehensive recommendation results. Extensive experiments on four datasets demonstrate that ContRec consistently outperforms both traditional and state-of-the-art LLM-based recommender systems. Our results highlight the potential of continuous tokenization and generative modeling for advancing the next generation of recommender systems.

Multi-Domain Recommendation (MDR) achieves the desirable recommendation performance by effectively utilizing the transfer information across different domains. Despite the great success, most existing MDR methods adopt a single structure to transfer complex domain-shared knowledge. However, the beneficial transferring information should vary across different domains. When there is knowledge conflict between domains or a domain is of poor quality, unselectively leveraging information from all domains will lead to a serious Negative Transfer Problem (NTP). Therefore, how to effectively model the complex transfer relationships between domains to avoid NTP is still a direction worth exploring. To address these issues, we propose a simple and dynamic Similar Domain Selection Principle (SDSP) for multi-domain recommendation in this paper. SDSP presents the initial exploration of selecting suitable domain knowledge for each domain to alleviate NTP. Specifically, we propose a novel prototype-based domain distance measure to effectively model the complexity relationship between domains. Thereafter, the proposed SDSP can dynamically find similar domains for each domain based on the supervised signals of the domain metrics and the unsupervised distance measure from the learned domain prototype. We emphasize that SDSP is a lightweight method that can be incorporated with existing MDR methods for better performance while not introducing excessive time overheads. To the best of our knowledge, it is the first solution that can explicitly measure domain-level gaps and dynamically select appropriate domains in the MDR field. Extensive experiments on three datasets demonstrate the effectiveness of our proposed method.

Nodes in the real-world graphs exhibit diverse patterns in numerous aspects, such as degree and homophily. However, most existent node predictors fail to capture a wide range of node patterns or to make predictions based on distinct node patterns, resulting in unsatisfactory classification performance. In this paper, we reveal that different node predictors are good at handling nodes with specific patterns and only apply one node predictor uniformly could lead to suboptimal result. To mitigate this gap, we propose a mixture of experts framework, MoE-NP, for node classification. Specifically, MoE-NP combines a mixture of node predictors and strategically selects models based on node patterns. Experimental results from a range of real-world datasets demonstrate significant performance improvements from MoE-NP.

Deep learning has been widely applied in recommender systems, which has recently achieved revolutionary progress. However, most existing learning-based methods assume that the user and item distributions remain unchanged between the training phase and the test phase. However, the distribution of user and item features can naturally shift in real-world scenarios, potentially resulting in a substantial decrease in recommendation performance. This phenomenon can be formulated as an Out-Of-Distribution (OOD) recommendation problem. To address this challenge, we propose a novelDualTest-Time-Training framework forOODRecommendation, termedDT3OR. In DT3OR, we incorporate a model adaptation mechanism during the test-time phase to carefully update the recommendation model, allowing the model to adapt specially to the shifting user and item features. To be specific, we propose a self-distillation task and a contrastive task to assist the model learning both the user’s invariant interest preferences and the variant user/item characteristics during the test-time phase, thus facilitating a smooth adaptation to the shifting features. Furthermore, we provide theoretical analysis to support the rationale behind our dual test-time training framework. To the best of our knowledge, this paper is the first work to address OOD recommendation via a test-time-training strategy. We conduct experiments on five datasets with various backbones. Comprehensive experimental results have demonstrated the effectiveness of DT3OR compared to other state-of-the-art baselines.

Recommender systems play a pivotal role across practical scenarios, showcasing remarkable capabilities in user preference modeling. However, the centralized learning paradigm predominantly used raises serious privacy concerns. The federated recommender system (FedRS) addresses this by updating models on clients, while a central server orchestrates training without accessing private data. Existing FedRS approaches, however, face unresolved challenges, including non-convex optimization, vulnerability, potential privacy leakage risk, and communication inefficiency. This paper addresses these challenges by reformulating the federated recommendation problem as a convex optimization issue, ensuring convergence to the global optimum. Based on this, we devise a novel method, RFRec, to tackle this optimization problem efficiently. In addition, we propose RFRecF, a highly efficient version that incorporates non-uniform stochastic gradient descent to improve communication efficiency. In user preference modeling, both methods learn local and global models, collaboratively learning users' common and personalized interests under the federated learning setting. Moreover, both methods significantly enhance communication efficiency, robustness, and privacy protection, with theoretical support. Comprehensive evaluations on four benchmark datasets demonstrate RFRec and RFRecF's superior performance compared to diverse baselines. The code is available to ease reproducibility1.

Graph Neural Networks (GNNs) have achieved promising performance in semi-supervised node classification in recent years. However, the problem of insufficient supervision, together with representation collapse, largely limits the performance of the GNNs in this field. To alleviate the collapse of node representations in semi-supervised scenario, we propose a novel graph contrastive learning method, termed M ixed G raph C ontrastive N etwork (MGCN). In our method, we improve the discriminative capability of the latent embeddings by an interpolation-based augmentation strategy and a correlation reduction mechanism. Specifically, we first conduct the interpolation-based augmentation in the latent space and then force the prediction model to change linearly between samples. Second, we enable the learned network to tell apart samples across two interpolation-perturbed views through forcing the correlation matrix across views to approximate an identity matrix. By combining the two settings, we extract rich supervision information from both the abundant unlabeled nodes and the rare yet valuable labeled nodes for discriminative representation learning. Extensive experimental results on six datasets demonstrate the effectiveness and the generality of MGCN compared to the existing state-of-the-art methods. The code of MGCN is available at https://github.com/xihongyang1999/MGCN on Github.

Graph neural networks (GNNs) have shown great potential in representation learning for various graph tasks. However, the distribution shift between the training and test sets poses a challenge to the efficiency of GNNs. To address this challenge, HomoTTT proposes a fully test-time training framework for GNNs to enhance the model’s generalization capabilities for node classification tasks. Specifically, our proposed HomoTTT designs a homophily-based and parameter-free graph contrastive learning task with adaptive augmentation to guide the model’s adaptation during the test-time training, allowing the model to adapt for specific target data. In the inference stage, HomoTTT proposes to integrate the original GNN model and the adapted model after TTT using a homophily-based model selection method, which prevents potential performance degradation caused by unconstrained model adaptation. Extensive experimental results on six benchmark datasets demonstrate the effectiveness of our proposed framework. Additionally, the exploratory study further validates the rationality of the homophily-based graph contrastive learning task with adaptive augmentation and the homophily-based model selection designed in HomoTTT .

With the prosperity of e-commerce and web applications, Recommender Systems (RecSys) have become an indispensable and important component in our daily lives, providing personalized suggestions that cater to user preferences. While Deep Neural Networks (DNNs) have achieved significant advancements in enhancing recommender systems by modeling user-item interactions and incorporating their textual side information, these DNN-based methods still exhibit some limitations, such as difficulties in effectively understanding users' interests and capturing textual side information, inabilities in generalizing to various seen/unseen recommendation scenarios and reasoning on their predictions, etc. Meanwhile, the development of Large Language Models (LLMs), such as ChatGPT and GPT-4, has revolutionized the fields of Natural Language Processing (NLP) and Artificial Intelligence (AI), due to their remarkable abilities in fundamental responsibilities of language understanding and generation, as well as impressive generalization capabilities and reasoning skills. As a result, recent studies have actively attempted to harness the power of LLMs to enhance recommender systems. Given the rapid evolution of this research direction in recommender systems, there is a pressing need for a systematic overview that summarizes existing LLM-empowered recommender systems, so as to provide researchers and practitioners in relevant fields with an in-depth understanding. Therefore, in this survey, we conduct a comprehensive review of LLM-empowered recommender systems from various aspects including pre-training, fine-tuning, and prompting paradigms. More specifically, we first introduce the representative methods to harness the power of LLMs (as a feature encoder) for learning representations of users and items. Then, we systematically review the emerging advanced techniques of LLMs for enhancing recommender systems from three paradigms, namely pre-training, fine-tuning, and prompting. Finally, we comprehensively discuss the promising future directions in this emerging field.

With the acceleration of urbanization, traffic forecasting has become an essential role in smart city construction. In the context of spatio-temporal prediction, the key lies in how to model the dependencies of sensors. However, existing works basically only consider the micro relationships between sensors, where the sensors are treated equally, and their macroscopic dependencies are neglected. In this paper, we argue to rethink the sensor's dependency modeling from two hierarchies: regional and global perspectives. Particularly, we merge original sensors with high intra-region correlation as a region node to preserve the inter-region dependency. Then, we generate representative and common spatio-temporal patterns as global nodes to reflect a global dependency between sensors and provide auxiliary information for spatio-temporal dependency learning. In pursuit of the generality and reality of node representations, we incorporate a Meta GCN to calibrate the regional and global nodes in the physical data space. Furthermore, we devise the cross-hierarchy graph convolution to propagate information from different hierarchies. In a nutshell, we propose a Hierarchical Information Enhanced Spatio-Temporal prediction method, HIEST, to create and utilize the regional dependency and common spatio-temporal patterns. Extensive experiments have verified the leading performance of our HIEST against state-of-the-art baselines. We publicize the code to ease reproducibility1.

In the era of information explosion, spatio-temporal data mining serves as a critical part of urban management. Considering the various fields demanding attention, e.g., traffic state, human activity, and social event, predicting multiple spatio-temporal attributes simultaneously can alleviate regulatory pressure and foster smart city construction. However, current research can not handle the spatio-temporal multi-attribute prediction well due to the complex relationships between diverse attributes. The key challenge lies in how to address the common spatio-temporal patterns while tackling their distinctions. In this paper, we propose an effective solution for spatio-temporal multi-attribute prediction, PromptST. We devise a spatio-temporal transformer and a parameter-sharing training scheme to address the common knowledge among different spatio-temporal attributes. Then, we elaborate a spatio-temporal prompt tuning strategy to fit the specific attributes in a lightweight manner. Through the pretrain and prompt tuning phases, our PromptST is able to enhance the specific spatio-temoral characteristic capture by prompting the backbone model to fit the specific target attribute while maintaining the learned common knowledge. Extensive experiments on real-world datasets verify that our PromptST attains state-of-the-art performance. Furthermore, we also prove PromptST owns good transferability on unseen spatio-temporal attributes, which brings promising application potential in urban computing. The implementation code is available to ease reproducibility.

With the rapid evolution of transformer architectures, researchers are exploring their application in sequential recommender systems (SRSs) and presenting promising performance on SRS tasks compared with former SRS models. However, most existing transformer-based SRS frameworks retain the vanilla attention mechanism, which calculates the attention scores between all item-item pairs. With this setting, redundant item interactions can harm the model performance and consume much computation time and memory. In this paper, we identify the sparse attention phenomenon in transformer-based SRS models and propose Sparse Transformer for sequential Recommendation tasks (STRec) to achieve the efficient computation and improved performance. Specifically, we replace self-attention with cross-attention, making the model concentrate on the most relevant item interactions. To determine these necessary interactions, we design a novel sampling strategy to detect relevant items based on temporal information. Extensive experimental results validate the effectiveness of STRec, which achieves the state-of-the-art accuracy while reducing 54% inference time and 70% memory cost. We also provide massive extended experiments to further investigate the property of our framework.

Deep recommender systems typically involve numerous feature fields for users and items, with a large number of low-frequency features. These low-frequency features would reduce the prediction accuracy with large storage space due to their vast quantity and inadequate training. Some pioneering studies have explored embedding compression techniques to address this issue of the trade-off between storage space and model predictability. However, these methods have difficulty compacting the embedding of low-frequency features in various feature fields due to the high demand for human experience and computing resources during hyper-parameter searching. In this paper, we propose the AutoDPQ framework, which automatically compacts low-frequency feature embeddings for each feature field to an adaptive magnitude. Experimental results indicate that AutoDPQ can significantly reduce the parameter space while improving recommendation accuracy. Moreover, AutoDPQ is compatible with various deep CTR models by improving their performance significantly with high efficiency.

Transformer models have achieved remarkable success in sequential recommender systems (SRSs). However, computing the attention matrix in traditional dot-product attention mechanisms results in a quadratic complexity with sequence lengths, leading to high computational costs for long-term sequential recommendation. Motivated by the above observation, we propose a novel L2-Normalized Linear Attention for the Transformer-based Sequential Recommender Systems (LinRec), which theoretically improves efficiency while preserving the learning capabilities of the traditional dot-product attention. Specifically, by thoroughly examining the equivalence conditions of efficient attention mechanisms, we show that LinRec possesses linear complexity while preserving the property of attention mechanisms. In addition, we reveal its latent efficiency properties by interpreting the proposed LinRec mechanism through a statistical lens. Extensive experiments are conducted based on two public benchmark datasets, demonstrating that the combination of LinRec and Transformer models achieves comparable or even superior performance than state-of-the-art Transformer-based SRS models while significantly improving time and memory efficiency. The implementation code is available online at https://github.com/Applied-Machine-Learning-Lab/LinRec.>

In the past few decades, artificial intelligence (AI) technology has experienced swift developments, changing everyone’s daily life and profoundly altering the course of human society. The intention behind developing AI was and is to benefit humans by reducing labor, increasing everyday conveniences, and promoting social good. However, recent research and AI applications indicate that AI can cause unintentional harm to humans by, for example, making unreliable decisions in safety-critical scenarios or undermining fairness by inadvertently discriminating against a group or groups. Consequently, trustworthy AI has recently garnered increased attention regarding the need to avoid the adverse effects that AI could bring to people, so people can fully trust and live in harmony with AI technologies. A tremendous amount of research on trustworthy AI has been conducted and witnessed in recent years. In this survey, we present a comprehensive appraisal of trustworthy AI from a computational perspective to help readers understand the latest technologies for achieving trustworthy AI. Trustworthy AI is a large and complex subject, involving various dimensions. In this work, we focus on six of the most crucial dimensions in achieving trustworthy AI: (i) Safety & Robustness, (ii) Nondiscrimination & Fairness, (iii) Explainability, (iv) Privacy, (v) Accountability & Auditability, and (vi) Environmental Well-being. For each dimension, we review the recent related technologies according to a taxonomy and summarize their applications in real-world systems. We also discuss the accordant and conflicting interactions among different dimensions and discuss potential aspects for trustworthy AI to investigate in the future.

Graph neural networks (GNNs) have been widely applied in the recommendation tasks and have achieved very appealing performance. However, most GNN-based recommendation methods suffer from the problem of data sparsity in practice. Meanwhile, pre-training techniques have achieved great success in mitigating data sparsity in various domains such as natural language processing (NLP) and computer vision (CV) . Thus, graph pre-training has the great potential to alleviate data sparsity in GNN-based recommendations. However, pre-training GNNs for recommendations faces unique challenges. For example, user-item interaction graphs in different recommendation tasks have distinct sets of users and items, and they often present different properties. Therefore, the successful mechanisms commonly used in NLP and CV to transfer knowledge from pre-training tasks to downstream tasks such as sharing learned embeddings or feature extractors are not directly applicable to existing GNN-based recommendations models. To tackle these challenges, we delicately design an adaptive graph pre-training framework for localized collaborative filtering (ADAPT) . It does not require transferring user/item embeddings, and is able to capture both the common knowledge across different graphs and the uniqueness for each graph simultaneously. Extensive experimental results have demonstrated the effectiveness and superiority of ADAPT.

Recent advances in multimodal single-cell technologies have enabled simultaneous acquisitions of multiple omics data from the same cell, providing deeper insights into cellular states and dynamics. However, it is challenging to learn the joint representations from the multimodal data, model the relationship between modalities, and, more importantly, incorporate the vast amount of single-modality datasets into the downstream analyses. To address these challenges and correspondingly facilitate multimodal single-cell data analyses, three key tasks have been introduced: Modality prediction, Modality matching andJoint embedding. In this work, we present a general Graph Neural Network framework scMoGNN to tackle these three tasks and show that scMoGNN demonstrates superior results in all three tasks compared with the state-of-the-art and conventional approaches. Our method is an official winner in the overall ranking ofModality prediction from NeurIPS 2021 Competition (https://openproblems.bio/neurips_2021/), and all implementations of our methods have been integrated into DANCE package (https://github.com/OmicsML/dance).

User-item interactions in recommendations can be naturally denoted as a user-item bipartite graph. Given the success of graph neural networks (GNNs) in graph representation learning, GNN-based Collaborative Filtering (CF) methods have been proposed to advance recommender systems. These methods often make recommendations based on the learned user and item embeddings. However, we found that they do not perform well with sparse user-item graphs which are quite common in real-world recommendations. Therefore, in this work, we introduce a novel perspective to build GNN-based CF methods for recommendations which leads to the proposed framework Localized Graph Collaborative Filtering (LGCF). One key advantage of LGCF is that it does not need to learn embeddings for each user and item, which is challenging in sparse scenarios. Alternatively, LGCF aims at encoding useful CF information into a localized graph and making recommendations based on such graph. Extensive experiments on various datasets validate the effectiveness of LGCF, especially in sparse scenarios. Furthermore, empirical results demonstrate that LGCF provides complementary information to the embedding-based CF model which can be utilized to boost recommendation performance.

Graphs such as social networks and molecular graphs are ubiquitous data structures in the real world. Due to their prevalence, it is of great research importance to extract meaningful patterns from graph structured data so that downstream tasks can be facilitated. Instead of designing hand-engineered features, graph representation learning has emerged to learn representations that can encode the abundant information about the graph. It has achieved tremendous success in various tasks such as node classification, link prediction, and graph classification and has attracted increasing attention in recent years.

Graph Neural Networks (GNNs) have achieved tremendous success in various real-world applications due to their strong ability in graph representation learning. GNNs explore the graph structure and node features by aggregating and transforming information within node neighborhoods. However, through theoretical and empirical analysis, we reveal that the aggregation process of GNNs tends to destroy node similarity in the original feature space. There are many scenarios where node similarity plays a crucial role. Thus, it has motivated the proposed framework SimP-GCN that can effectively and efficiently preserve node similarity while exploiting graph structure. Specifically, to balance information from graph structure and node features, we propose a feature similarity preserving aggregation which adaptively integrates graph structure and node features. Furthermore, we employ self-supervised learning to explicitly capture the complex feature similarity and dissimilarity relations between nodes. We validate the effectiveness of SimP-GCN on seven benchmark datasets including three assortative and four disassorative graphs. The results demonstrate that SimP-GCN outperforms representative baselines. Further probe shows various advantages of the proposed framework. The implementation of SimP-GCN is available at https://github.com/ChandlerBang/SimP-GCN.

Graph Convolutional Networks (GCNs) show promising results for semi-supervised learning tasks on graphs, thus become favorable comparing with other approaches. Despite the remarkable success of GCNs, it is difficult to train GCNs with insufficient supervision. When labeled data are limited, the performance of GCNs becomes unsatisfying for low-degree nodes. While some prior work analyze successes and failures of GCNs on the entire model level, profiling GCNs on individual node level is still underexplored.

Graph Neural Networks (GNNs), which extend deep neural networks to graph-structured data, have attracted increasing attention. They have been proven to be powerful for numerous graph related tasks such as graph classification, link prediction, and node classification. To adapt GNNs to graph classification, recent works aim to learn graph-level representation through a hierarchical pooling procedure. One major direction is to select important nodes to hierarchically coarsen the input graph and gradually reduce the information into the graph representation. However, most of the existing methods only select important nodes, which can be redundant and cannot represent the original graph well. Meanwhile, the information of non-selected nodes is often overlooked when generating a new coarser graph, which may lead to the tremendous loss of important structural and node feature information. In this paper, we propose a novel pooling operator RepPool to learn hierarchical graph representations. Specifically, we introduce the concept of representativeness that is combined with the importance for node selection and we provide a learnable way to integrate non-selected nodes. By combining the RepPool operator with conventional GCN convolutional layers, a hierarchical graph classification architecture is developed. Extensive experiments on various public benchmarks have demonstrated the effectiveness of the proposed method. The implementation of the proposed framework is available.

Many real data come in the form of non-grid objects, i.e. graphs, from social networks to molecules. Adaptation of deep learning from grid-alike data (e.g. images) to graphs has recently received unprecedented attention from both machine learning and data mining communities, leading to a new cross-domain field---Deep Graph Learning (DGL). Instead of painstaking feature engineering, DGL aims to learn informative representations of graphs in an end-to-end manner. It has exhibited remarkable success in various tasks, such as node/graph classification, link prediction, etc.

Traffic flow analysis, prediction and management are keystones for building smart cities in the new era. With the help of deep neural networks and big traffic data, we can better understand the latent patterns hidden in the complex transportation networks. The dynamic of the traffic flow on one road not only depends on the sequential patterns in the temporal dimension but also relies on other roads in the spatial dimension. Although there are existing works on predicting the future traffic flow, the majority of them have certain limitations on modeling spatial and temporal dependencies. In this paper, we propose a novel spatial temporal graph neural network for traffic flow prediction, which can comprehensively capture spatial and temporal patterns. In particular, the framework offers a learnable positional attention mechanism to effectively aggregate information from adjacent roads. Meanwhile, it provides a sequential component to model the traffic flow dynamics which can exploit both local and global temporal dependencies. Experimental results on various real traffic datasets demonstrate the effectiveness of the proposed framework.