论文检索

Shenghao Yang 0004, Weizhi Ma, Zhiqiang Guo, Min Zhang 0006, Haiyang Wu, Junjie Zhai, Chunhui Zhang, Yuekui Yang

With the development of multi-modal modeling techniques, recent sequential recommender systems enhance transferability by incorporating cross-domain universal multi-modal data, e.g., text and image. Existing methods typically adopt pairwise alignment to alleviate the gap between modalities. However, this alignment paradigm has limitations on explainability, consistency, and expansibility, resulting in suboptimal performance. This paper proposes a novel Explainable multi-modality Alignment method for transferable Rec ommender systems, i.e., EARec. Specifically, we design a two-stage framework to achieve explainable modality alignment in the source domain and recommendation based on aligned modality representations in the target domain. In the first stage, we adopt a generative task to align various modalities in parallel to a shared anchor with explainable meaning. All modalities share the same anchor to ensure consistent direction. Additionally, we treat behavior as an independent modality to integrate task-specific information into the alignment framework. In the second stage, we compose multiple item modality representation models trained in the first stage to obtain a unified model capable of understanding various modalities simultaneously, thereby providing high-quality item modality representations for recommendations in the target domain. Benefiting from the approach of parallel modality alignment followed by model composition, the framework shows flexibility in expanding new modalities. Experimental results on multiple public datasets demonstrate the superiority of EARec over baselines, and further analyses indicate the explainability and expansibility of the proposed alignment method.

Xin Wang 0117, Ling Feng, Huijun Zhang, Lei Cao, Kaisheng Zeng, Qi Li 0051, Yang Ding 0003, Yi Dai, David A. Clifton

Stress haunts people in modern society, which may cause severe health issues if left unattended. With social media becoming an integral part of daily life, leveraging social media to detect stress has gained increasing attention. While the majority of the work focuses on classifying stress states and stress categories, this study introduce a new task aimed at estimating more specific stressors (like exam, writing paper, etc.) through users' posts on social media. Unfortunately, the diversity of stressors with many different classes but a few examples per class, combined with the consistent arising of new stressors over time, hinders the machine understanding of stressors. To this end, we cast the stressor estimation problem within a practical scenario few-shot learning setting, and propose a novel meta-learning based stressor estimation framework that is enhanced by a meta-knowledge inheritance mechanism. This model can not only learn generic stressor context through meta-learning, but also has a good generalization ability to estimate new stressors with little labeled data. A fundamental breakthrough in our approach lies in the inclusion of the meta-knowledge inheritance mechanism, which equips our model with the ability to prevent catastrophic forgetting when adapting to new stressors. The experimental results show that our model achieves state-of-the-art performance compared with the baselines. Additionally, we construct a social media-based stressor estimation dataset that can help train artificial intelligence models to facilitate human well-being.

Xiaoshuai Song, Zhengyang Wang, Keqing He 0001, Guanting Dong 0001, Yutao Mou, Jinxu Zhao, Weiran Xu

The rapid evolution of the Web as a key platform for information dissemination has led to the growing integration of large language models (LLMs) in Web-based applications. However, the swift changes in web content present challenges in maintaining these models' relevance and accuracy. The task of Knowledge Editing (KE) is aimed at efficiently and precisely adjusting the behavior of large language models (LLMs) to update specific knowledge while minimizing any adverse effects on other knowledge. Current research predominantly concentrates on editing white-box LLMs, neglecting a significant scenario: editing black-box LLMs, where access is limited to interfaces and only textual output is provided. In this paper, we initially officially introduce KE on black-box LLMs, followed by presenting a thorough evaluation framework. This framework operates without requiring logits and considers pre- and post-edit consistency, addressing the limitations of current evaluations that are inadequate for black-box LLMs editing and lack comprehensiveness. To address privacy leaks of editing data and style over-editing in existing approaches, we propose a new postEdit framework. postEdit incorporates a retrieval mechanism for editing knowledge and a purpose-trained editing plugin called post-editor, ensuring privacy through downstream processing and maintaining textual style consistency via fine-grained editing. Experiments and analysis conducted on two benchmarks show that postEdit surpasses all baselines and exhibits robust generalization, notably enhancing style retention by an average of +20.82%. Our code is available on github https://github.com/songxiaoshuai/postEdit.

Tao Tan 0005, Yining Qian, Ang Lv, Hongzhan Lin 0002, Songhao Wu, Yongbo Wang, Feng Wang 0023, Jingtong Wu, Xin Lu, Rui Yan 0001

Large language models (LLMs) enhanced with retrieval-augmented generation (RAG) have introduced a new paradigm for web search. However, the limited context awareness of LLMs degrades their performance on RAG tasks. Existing methods to enhance context awareness are often inefficient, incurring time or memory overhead during inference, and many are tailored to specific position embeddings. In this paper, we propose Position-Embedding-Agnostic attention Re-weighting (PEAR), which enhances the context awareness of LLMs with zero inference overhead. Specifically, on a proxy task focused on context copying, we first detect heads which suppress the models' context awareness, thereby diminishing RAG performance. To weaken the impact of these heads, we re-weight their outputs with learnable coefficients. The LLM (with frozen parameters) is optimized by adjusting these coefficients to minimize loss on the proxy task. During inference, the optimized coefficients are fixed to re-weight these heads, regardless of the specific task at hand. Our proposed PEAR offers two major advantages over previous approaches: (1) It introduces zero additional inference overhead in terms of memory usage or inference time, while outperforming competitive baselines in accuracy and efficiency across various RAG tasks. (2) It is independent of position embedding algorithms, ensuring broader applicability. Our code is available at https://github.com/TTArch/PEAR-RAG.

Bailu Ding, Jiaqi Zhai

Retrieval plays a fundamental role in recommendation systems, search, and natural language processing (NLP) by efficiently finding relevant items from a large corpus given a query. Dot products have been widely used as the similarity function in such tasks, enabled by Maximum Inner Product Search (MIPS) algorithms for efficient retrieval. However, state-of-the-art retrieval algorithms have migrated to learned similarities. These advanced approaches encompass multiple query embeddings, complex neural networks, direct item ID decoding via beam search, and hybrid solutions. Unfortunately, we lack efficient solutions for retrieval in these state-of-the-art setups. Our work addresses this gap by investigating efficient retrieval techniques with expressive learned similarity functions. We establish Mixture-of-Logits (MoL) as a universal approximator of similarity functions, demonstrate that MoL's expressiveness can be realized empirically to achieve superior performance on diverse retrieval scenarios, and propose techniques to retrieve the approximate top-k results using MoL with tight error bounds. Through extensive experimentation, we show that MoL, enhanced by our proposed mutual information-based load balancing loss, sets new state-of-the-art results across heterogeneous scenarios, including sequential retrieval models in recommendation systems and finetuning language models for question answering; and our approximate top-k algorithms outperform baselines by up to 66× in latency while achieving >.99 recall rate compared to exact algorithms.

Theresia Veronika Rampisela, Tuukka Ruotsalo, Maria Maistro, Christina Lioma

Fairness and relevance are two important aspects of recommender systems (RSs). Typically, they are evaluated either (i) separately by individual measures of fairness and relevance, or (ii) jointly using a single measure that accounts for fairness with respect to relevance. However, approach (i) often does not provide a reliable joint estimate of the goodness of the models, as it has two different best models: one for fairness and another for relevance. Approach (ii) is also problematic because these measures tend to be ad-hoc and do not relate well to traditional relevance measures, like NDCG. Motivated by this, we present a new approach for jointly evaluating fairness and relevance in RSs: Distance to Pareto Frontier (DPFR). Given some user-item interaction data, we compute their Pareto frontier for a pair of existing relevance and fairness measures, and then use the distance from the frontier as a measure of the jointly achievable fairness and relevance. Our approach is modular and intuitive as it can be computed with existing measures. Experiments with 4 RS models, 3 re-ranking strategies, and 6 datasets show that existing metrics have inconsistent associations with our Pareto-optimal solution, making DPFR a more robust and theoretically well-founded joint measure for assessing fairness and relevance. Our code: https://github.com/theresiavr/DPFR-recsys-evaluation

Yifan Wang, Gonzalo R. Arce, Guangmo Tong

Hypergraph neural networks have been promising tools for handling learning tasks involving higher-order data, with notable applications in web graphs, such as modeling multi-way hyperlink structures and complex user interactions. Yet, their generalization abilities in theory are less clear to us. In this paper, we seek to develop margin-based generalization bounds for four representative classes of hypergraph neural networks, including convolutional-based methods (UniGCN), set-based aggregation (AllDeepSets), invariant and equivariant transformations (M-IGN), and tensor-based approaches (T-MPHN). Through the PAC-Bayes framework, our results reveal the manner in which hypergraph structure and spectral norms of the learned weights can affect the generalization bounds, where the key technical challenge lies in developing new perturbation analysis for hypergraph neural networks, which offers a rigorous understanding of how variations in the model's weights and hypergraph structure impact its generalization behavior. Our empirical study examines the relationship between the practical performance and theoretical bounds of the models over synthetic and real-world datasets. One of our primary observations is the strong correlation between the theoretical bounds and empirical loss, with statistically significant consistency in most cases.

Dongxiao He, Yongqi Huang, Jitao Zhao, Xiaobao Wang, Zhen Wang 0004

Graph Contrastive Learning (GCL) is a widely adopted approach in self-supervised graph representation learning, applying contrastive objectives to produce effective representations. However, current GCL methods primarily focus on capturing implicit semantic relationships, often overlooking the structural commonsense embedded within the graph's structure and attributes, which contains underlying knowledge crucial for effective representation learning. Due to the lack of explicit information and clear guidance in general graph, identifying and integrating such structural commonsense in GCL poses a significant challenge. To address this gap, we propose a novel framework called Structural Commonsense Unveiling in Graph Contrastive Learning (Str-GCL). Str-GCL leverages first-order logic rules to represent structural commonsense and explicitly integrates them into the GCL framework. It introduces topological and attribute-based rules without altering the original graph and employs a representation alignment mechanism to guide the encoder in effectively capturing this commonsense. To the best of our knowledge, this is the first attempt to directly incorporate structural commonsense into GCL. Extensive experiments demonstrate that Str-GCL outperforms existing GCL methods, providing a new perspective on leveraging structural commonsense in graph representation learning.

Wenwen Meng, Chuan Ma 0001, Ming Ding 0001, Chunpeng Ge 0001, Yuwen Qian, Tao Xiang 0001

Website Fingerprinting (WF) attacks allow passive eavesdroppers to deduce the websites a user visits by analyzing encrypted traffic, threatening user privacy. While current WF attacks achieve high accuracy, they typically assume single-tab browsing, which is unrealistic as users often open multiple tabs, creating mixed traffic. Existing multi-tab WF approaches require large datasets and frequent retraining due to evolving website content, limiting their practicality. In this paper, we introduce Few-shot Multi-tab Website Fingerprinting (FMWF), a novel approach designed to address the limitations of existing multi-tab WF attacks. FMWF directly tackles the challenges of mixed, overlapping traffic traces generated from multi-tab browsing, leveraging two key innovations: (1) an advanced data augmentation technique that synthesizes realistic multi-tab traffic sequences from easily collected single-tab traces, thereby dramatically reducing the need for large-scale real-world traffic data; and (2) a powerful fine-tuning algorithm based on transfer learning that adapts pre-trained models to new, multi-tab environments with minimal additional data. This two-stage framework enables FMWF to capture the complex effectively, overlapping traffic patterns inherent in multi-tab browsing while maintaining a high level of flexibility and significantly lowering computational and data collection burdens. Our experiments, conducted using real traffic traces collected from three widely-used browsers-Microsoft Edge, Google Chrome, and Tor Browser-highlight the superior performance of FMWF in both closed-world and open-world scenarios. Notably, FMWF achieves a minimum 12.3% improvement in accuracy compared to ARES (SP'23) [7], TMWF (CCS'23) [13], and BAPM (ACSAC'21) [10] in the open-world scenario. The code with related datasets is available at https://github.com/WW-Meng/FMWF.

Kyungchan Lim, Kiho Lee, Fujiao Ji, Yonghwi Kwon 0001, Hyoungshick Kim, Doowon Kim

Phishing attacks pose a significant threat to Internet users. Understanding the security posture of phishing infrastructure is crucial for developing effective defense strategies, as it helps identify potential weaknesses that attackers might exploit. Despite extensive research, there may still be a gap in fully understanding these security weaknesses. To address this important issue, this paper presents a longitudinal study of security configurations and vulnerabilities in phishing websites and associated kits. We focus on two main areas: (1) analyzing the security configurations of phishing websites and servers, particularly HTTP headers and application-level security, and (2) examining the prevalence and types of vulnerabilities in phishing kits. We analyze data from 906,731 distinct phishing websites collected over 2.5 years, covering HTML headers, client-side resources, and phishing kits. Our findings suggest that phishing websites often employ weak security configurations, with 88.8% of the 13,344 collected phishing kits containing at least one potential vulnerability, and 12.5% containing backdoor vulnerabilities. These vulnerabilities present an opportunity for defenders to shift from passive defense to active disruption of phishing operations. Our research proposes a new approach to leverage weaknesses in phishing infrastructure, allowing defenders to take proactive actions to disable phishing sites earlier and reduce their effectiveness.

Jiarui Qin, Weiwen Liu, Weinan Zhang 0001, Yong Yu 0001

A vast amount of user behavior data is constantly accumulating on today's large recommendation platforms, recording users' various interests and tastes. Preserving knowledge from the old data while new data continually arrives is a vital problem for recommender systems. Existing approaches generally seek to save the knowledge implicitly in the model parameters. However, such a parameter-centric approach lacks scalability and flexibility---the capacity is hard to scale, and the knowledge is inflexible to utilize. Hence, in this work, we propose a framework that turns massive user behavior data to retrievable knowledge (D2K). It is a data-centric approach that is model-agnostic and easy to scale up. Different from only storing unary knowledge such as the user-side or item-side information, D2K propose to store ternary knowledge for recommendation, which is determined by the complete recommendation factors---user, item, and context. The knowledge retrieved by target samples can be directly used to enhance the performance of any recommendation algorithms. Specifically, we introduce a Transformer-based knowledge encoder to transform the old data into knowledge with the user-item-context cross features. A personalized knowledge adaptation unit is devised to effectively exploit the information from the knowledge base by adapting the retrieved knowledge to the target samples. Extensive experiments demonstrate the superiority of the proposed method.

Cheng Wang 0025, Wenchao Xu 0001, Haozhao Wang, Wei Liu 0144, Ruixuan Li 0001

Privacy-preserving Cross-Domain Recommendation (CDR) has been extensively studied to address the cold-start problem using auxiliary source domains while simultaneously protecting sensitive information. However, existing privacy-preserving CDR methods rely heavily on transferring sensitive user embeddings or behaviour logs, which leads to adopt privacy methods to distort the data patterns before transferring it to the target domain. The distorted information can compromise overall performance during the knowledge transfer process. To overcome these challenges, our approach differs from existing privacy-preserving methods that focus on safeguarding user-sensitive information. Instead, we concentrate on distilling transferable knowledge from insensitive item embeddings, which we refer to as prototypes. Specifically, we propose a conditional model inversion mechanism to accurately distill prototypes for individual users. We have designed a new data format and corresponding learning paradigm for distilling transferable prototypes from traditional recommendation models using model inversion. These prototypes facilitate bridging the domain shift between distinct source and target domains in a privacy-friendly manner. Additionally, they enable the identification of top-k users in the target domain to substitute for cold-start users prediction. We conduct extensive experiments across large real-world datasets, and the results substantiate the effectiveness of PFCDR https://github.com/walcheng/PFCDR..

Zhenghai Xue, Qingpeng Cai 0001, Bin Yang 0042, Lantao Hu, Peng Jiang 0002, Kun Gai, Bo An 0001

The field of Reinforcement Learning (RL) has garnered increasing attention for its ability of optimizing user retention in recommender systems. A primary obstacle in this optimization process is the environment non-stationarity stemming from the continual and complex evolution of user behavior patterns over time, such as variations in interaction rates and retention propensities. These changes pose significant challenges to existing RL algorithms for recommendations, leading to issues with dynamics and reward distribution shifts. This paper introduces a novel approach called Adaptive User Retention Optimization (AURO) to address this challenge. To navigate the recommendation policy in non-stationary environments, AURO introduces an state abstraction module in the policy network. The module is trained with a new value-based loss function, aligning its output with the estimated performance of the current policy. As the policy performance of RL is sensitive to environment drifts, the loss function enables the state abstraction to be reflective of environment changes and notify the recommendation policy to adapt accordingly. Additionally, the non-stationarity of the environment introduces the problem of implicit cold start, where the recommendation policy continuously interacts with users displaying novel behavior patterns. AURO encourages exploration guarded by performance-based rejection sampling to maintain a stable recommendation quality in the cost-sensitive online environment. Extensive empirical analysis are conducted in a user retention simulator, the MovieLens dataset, and a live short-video recommendation platform, demonstrating AURO's superior performance against all evaluated baseline algorithms. Code is available at https://github.com/AIDefender/AURO

Huanyu Zhang 0001, Xiaoxuan Shen, Baolin Yi, Jianfang Liu, Yinao Xie

Incorporating a critiquing component into recommender applications facilitates the enhancement of user perception. Typically, critique-able recommender systems adapt the model parameters and update the recommendation list in real-time through the analysis of user critiquing keyphrases in the inference phase. The current critiquing methods necessitate the designation of a dedicated recommendation model to estimate user relevance to the critiquing keyphrase during the training phase preceding the recommendations update. This paradigm restricts the applicable scenarios and reduces the potential for keyphrase exploitation. Furthermore, these approaches ignore the issue of catastrophic forgetting caused by continuous modification of model parameters in multi-step critiquing. Thus, we present a general Representative Items Sampling Framework for Critiquing on Knowledge Graph Recommendation (RISC) implemented as a plug-in, which offers a new paradigm for critiquing in mainstream recommendation scenarios. RISC leverages the knowledge graph to sample important representative items as a hinge to expand and convey information from user critiquing, indirectly estimating the relevance of the user to the critiquing keyphrase. Consequently, the necessity for specialized user-keyphrase correlation modules is eliminated with respect to a variety of knowledge graph recommendation models. Moreover, we propose a Weight Experience Replay (WER) approach based on KG to mitigate catastrophic forgetting by reinforcing the user's prior preferences during the inference phase. Our extensive experimental findings on three real-world datasets and three knowledge graph recommendation methods illustrate that RISC with WER can be effectively integrated into knowledge graph recommendation models to efficiently utilize user critiquing for refining recommendations and mitigate catastrophic forgetting.

Songwen Hu, Ryan A. Rossi, Tong Yu 0001, Junda Wu, Handong Zhao, Sungchul Kim, Shuai Li 0010

Visualization recommendation aims to enable rapid visual analysis of massive datasets. In real-world scenarios, it is essential to quickly gather and comprehend user preferences to cover users from diverse backgrounds, including varying skill levels and analytical tasks. Previous approaches to personalized visualization recommendations are non-interactive and rely on initial user data for new users. As a result, these models cannot effectively explore options or adapt to real-time feedback. To address this limitation, we propose an interactive personalized visualization recommendation (PVisRec) system that learns on user feedback from previous interactions. For more interactive and accurate recommendations, we propose Hier-SUCB, a contextual combinatorial semi-bandit in the PVisRec setting. Theoretically, we show an improved overall regret bound with the same rank of time but an improved rank of action space. We further demonstrate the effectiveness of Hier-SUCB through extensive experiments where it is comparable to offline methods and outperforms other bandit algorithms in the setting of visualization recommendation.

Jiamin Luo, Jingjing Wang, Junxiao Ma, Yujie Jin, Shoushan Li, Guodong Zhou 0001

Prior studies on Visual Sentiment Understanding (VSU) primarily rely on the explicit scene information (e.g., facial expression) to judge visual sentiments, which largely ignore implicit scene information (e.g., human action, objection relation and visual background), while such information is critical for precisely discovering visual sentiments. Motivated by this, this paper proposes a new Omni-scene driven visual Sentiment Identifying, Locating and Attributing in videos (Omni-SILA) task, aiming to interactively and precisely identify, locate and attribute visual sentiments through both explicit and implicit scene information. Furthermore, this paper believes that this Omni-SILA task faces two key challenges: modeling scene and highlighting implicit scene beyond explicit. To this end, this paper proposes an Implicit-enhanced Causal MoE (ICM) approach for addressing the Omni-SILA task. Specifically, a Scene-Balanced MoE (SBM) and an Implicit-Enhanced Causal (IEC) blocks are tailored to model scene information and highlight the implicit scene information beyond explicit, respectively. Extensive experimental results on our constructed explicit and implicit Omni-SILA datasets demonstrate the great advantage of the proposed ICM approach over advanced Video-LLMs.

Aidong Zhang 0001, Vipin Kumar 0001, Yan Liu 0002

The past decade has been an inspiring time for artificial intelligence (AI) research. AI systems have transformed norms and practices across industries and have permeated the fabric of human society. Moreover, AI is ushering in a transformative technological age by making remarkable breakthroughs in a number of scientific fields such as protein structure prediction and medical imaging. There is increasing consensus in the wider scientific community that AI is poised to disrupt science by unlocking entirely new approaches, driving new scientific inquiry, and enabling greater scientific leaps with far-reaching social consequences. However, there are substantial barriers preventing science from realizing that potential, and addressing these barriers will require support for advances in AI methods and the adoption of these methods in routine scientific research. In this special day at KDD 2025, we host a series of talks by distinguished researchers on AI for science.

Kun Yi 0001, Qi Zhang 0020, Wei Fan 0010, Longbing Cao, Shoujin Wang, Hui He, Guodong Long, Liang Hu 0004, Qingsong Wen, Hui Xiong 0001

Recently, frequency transformation (FT) has been increasingly incorporated into deep learning models to significantly enhance state-of-the-art accuracy and efficiency in time series analysis. The advantages of FT, such as high efficiency and a global view, have been rapidly explored and exploited in various time series tasks and applications, demonstrating the promising potential of FT as a new deep learning paradigm for time series analysis. Despite the growing attention and the proliferation of research in this emerging field, there is currently a lack of a systematic review and in-depth analysis of deep learning-based time series models with FT. It is also unclear why FT can enhance time series analysis and what its limitations are in the field. To address these gaps, we present a comprehensive review that systematically investigates and summarizes the recent research advancements in deep learning-based time series analysis with FT. Specifically, we explore the primary approaches used in current models that incorporate FT, the types of neural networks that leverage FT, and the representative FT-equipped models in deep time series analysis. We propose a novel taxonomy to categorize the existing methods in this field, providing a structured overview of the diverse approaches employed in incorporating FT into deep learning models for time series analysis. Finally, we highlight the advantages and limitations of FT for time series modeling and identify potential future research directions that can further contribute to the community of time series analysis.

Xiaoou Liu, Tiejin Chen, Longchao Da, Chacha Chen, Zhen Lin 0001, Hua Wei 0001

Uncertainty quantification (UQ) enhances the reliability of Large Language Models (LLMs) by estimating confidence in outputs, enabling risk mitigation and selective prediction. However, traditional UQ methods struggle with LLMs due to computational constraints and decoding inconsistencies. Moreover, LLMs introduce unique uncertainty sources, such as input ambiguity, reasoning path divergence, and decoding stochasticity, that extend beyond classical aleatoric and epistemic uncertainty. To address this, we introduce a new taxonomy that categorizes UQ methods based on computational efficiency and uncertainty dimensions, including input, reasoning, parameter, and prediction uncertainty. We evaluate existing techniques, summarize existing benchmarks and metrics for UQ, assess their real-world applicability, and identify open challenges, emphasizing the need for scalable, interpretable, and robust UQ approaches to enhance LLM reliability.

Clarice Wang, Yimin Shi 0001, Xiaokui Xiao

Traditional evaluations of conversational AI have primarily focused on ''closed-ended'' interactions, where a human user queries the AI system, such as in customer support. However, many advanced real-world applications-such as job interviews, podcast hosting, and legal or healthcare intake discussions-require ''open-ended'' interactions in which the AI must take initiative by formulating questions to fully understand the human user's story. As AI assumes broader roles that demand greater autonomy, evaluating its performance in open-ended conversations becomes significantly more complex. This paper introduces a novel framework for rigorously assessing AI agents in such open-ended interviews. In this framework, a secondary AI agent (Agent B) simulates a human interviewee by strictly following a structured script that defines topic strengths and weaknesses, along with guidelines for when to hint at or deviate from these topics. Meanwhile, the primary AI agent (Agent A) engages in dynamic questioning to uncover the script's underlying facts and narrative. By comparing the final conversation transcript to the original script, we assess Agent A using metrics such as completeness, consistency, and investigative depth. This approach not only establishes a new benchmark for open-ended conversational skills but also provides insight into how effectively an AI agent can detect and navigate strategic diversions in scripted behavior.