paperlist.ai

AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

IJCAI 2026

FedCIGAR: A Personalized Reconstruction Approach for Federated Graph-Level Anomaly Detection

Yunfeng Zhao, Yixin Liu, Qingfeng Chen, Shiyuan Li, Yue Tan, Shirui Pan

Graph-level anomaly detection (GLAD) is crucial for ensuring the reliability of graph-driven applications by identifying abnormal graphs that deviate from the majority. Considering the privacy concerns in distributed scenarios, federated graph-level anomaly detection (FedGLAD) has emerged as a promising solution to enable collaborative detection without sharing raw data. However, existing methods suffer from poor generalization due to the reliance on unrealistic synthetic anomalies and insufficient personalization capabilities under data heterogeneity. To address these challenges, we propose a novel Federated graph-level anomaly detection approach with Cluster-adaptIve GAted Reconstruction (FedCIGAR). Specifically, we design a reconstruction-based paradigm trained on normal graphs to avoid synthetic data. Furthermore, we introduce a client-side node contribution gating mechanism and a server-side sliding window-based clustering strategy to tackle data heterogeneity. Extensive experiments demonstrate that FedCIGAR achieves superior performance and robustness compared to state-of-the-art methods.

IJCAI 2026

Graph Anomaly Detection via Feature Selection with Local Topological Residuals

Yazheng Zhao, Nannan Wu, haoran yin, Yiming Zhao

Graph anomaly detection (GAD) aims to identify nodes that exhibit significant deviations from expected structural or attribute patterns, and has garnered increasing attention in recent years. Recent approaches for GAD have predominantly focused on local inconsistency mining, which refers to the difficulty of establishing high similarity relationships between anomalous nodes and their neighbors. While local inconsistency mining requires the incorporation of topological information, the use of Graph Neural Networks (GNNs) to introduce such information tends to homogenize connected nodes, thereby causing the loss of local anomalous signals. To address this challenge, we propose LTRGAD, a two-stage GAD framework that performs feature selection based on local feature-topological residuals (LTR). By processing features separately, LTRGAD effectively introduces topological information while preserving the original local anomalous patterns, enabling more accurate local anomaly detection. Subsequently, global anomaly detection is conducted on the entire graph by leveraging the results from the local detection phase. Extensive experiments on seven benchmark datasets demonstrate the effectiveness of the proposed LTRGAD framework.

IJCAI 2026

Pivot-Centric Trajectory Prediction: Bridging Long Horizons via Dynamical Guidance

Xiucong Zhao, Jindong Tian, Hao Miao

Forecasting precise future motion of surrounding agents is essential for reliable autonomous vehicles. However, as the demand for longer prediction horizons increases, existing endpoint-completion or iterative-refine methods increasingly struggle with weak guidance and compounding errors. To tackle the long-horizon prediction challenge, we propose Pivot-Centric Trajectory Prediction (PCTP). By introducing "pivots" and focusing on predicting pivot points along extended trajectories, we divide the long-term prediction task into short-term sub-tasks at various scales. Specifically, PCTP decouples the long-term trajectory predicting process into two processes: pivot prediction and pivot-based trajectory refinement. The pivot prediction process aims to utilize global map context and agent-to-agent interactions to identify these "pivot points", while the pivot-based trajectory refinement process focuses on local map details and refines the short-term trajectory based on predicted "pivot points". Compared with existing methods, PCTP provides more intermediate guidance while reducing compounding errors. Moreover, PCTP is a flexible approach that can be integrated into most state-of-the-art trajectory prediction models. Experimental results show that PCTP improves the prediction accuracy of leading models on both Argoverse I and Argoverse II datasets with minimal impact on model size. Specifically, PCTP combined with QCNet outperforms all published ensemble-free methods on the Argoverse II leaderboard at submission.

IJCAI 2026

DaV-Gen: End-to-End Generative Retrieval via Draft-and-Verify

Meng Zhao, Chunmei Liu, Qinyong Wang

Mainstream industrial information retrieval systems (e.g., search and recommendation) are usually built upon Multi-Stage Cascade Architectures (MCAs), which balance effectiveness and efficiency through a coarse-to-fine "retrieval-ranking" pipeline. However, the optimization objectives across different stages are substantially inconsistent, propagating or even amplifying the early-stage errors that ultimately degrade the quality of final results. While emerging end-to-end generative models offer a potential solution by unifying the pipeline, their online serving performance is severely hindered by the auto-regressive process inherited from the standard decoder-only structure. To bridge this gap, we introduce DaV-Gen, a novel unified solution designed to fundamentally refactor the paradigm for both search and recommendation via a "Draft-and-Verify" mechanism. Inspired by the process used by speculative decoding, our framework redesigns the generation task into two synergistic operations within a single model. During training, the model is concurrently optimized for both candidate drafting and fine-grained verification. This is achieved by a composite loss function that jointly trains the model on two distinct but related objectives: 1) a contrastive loss that structures the embedding space for efficient drafting, and 2) a fusion loss that combines generative likelihood with vector similarity to produce a superior verification score. This integrated training strategy equips the model with dual capabilities. At inference time, it first performs highly efficient vector-based drafting to generate a candidate set, and then verifies these candidates using the more powerful fused scoring function, thereby achieving both the speed of sparse drafting and the precision of advanced generative models within a unified, end-to-end architecture.

IJCAI 2026

DynaOD: Dynamic Origin-Destination Flow Generation with Discrete-to-Continuous Temporal Semantic Modeling

Jie Zhao, Xianqi Dai, Jie Feng, Huandong Wang, Yong Li

Dynamic origin-destination (OD) flow generation seeks to synthesize realistic mobility dynamics from temporal context alone, without relying on historical OD observations. A key challenge is to translate semantic temporal signals into temporally coherent OD patterns while preserving the inherent spatial heterogeneity of urban regions. We propose DynaOD, a semantic-driven framework that models temporal dynamics through two complementary perspectives: discrete directional trends that characterize qualitative shifts in urban activity patterns, and continuous temporal evolution that captures how such shifts unfold over time. By jointly encoding these temporal semantics, the framework constructs time-varying region representations that condition pretrained static OD generators in a lightweight and plug-and-play fashion. This modular design further supports scalable deployment and cross-city transferability. Extensive experiments on large-scale real-world datasets show that our method consistently outperforms representative baselines in both predictive accuracy and distributional fidelity. Code is publicly available at https://github.com/csjiezhao/DynaOD.

IJCAI 2026

Geodesic Expert Routing for Unbiased Knowledge Distillation in Recommendation

Xuan Zhang, Rongchuan Wei, Chunyu Wei, Hongxing Yuan, Yushun Fan

Knowledge distillation has become a prevalent technique for deploying efficient recommender systems, enabling lightweight student models to approximate the performance of larger teachers. However, we identify a critical issue: distillation systematically amplifies popularity bias, as student models inherit and intensify the popularity-driven shortcuts encoded in teachers trained on interaction data dominated by popular items. To address this limitation, we propose GUIDE (Geodesic aware Unbiased Instructive Distillation with Experts), a collaborative distillation framework that incorporates domain-specific debiasing experts alongside the global teacher. GUIDE tackles two key challenges in this paradigm. First, for expert routing, we introduce Spherical Expert Alignment, which conducts expert-student matching on the spherical manifold with geodesic distance optimization, eliminating magnitude-induced bias and ensuring stable gradient flow. Second, for context fusion, we design a Meta-Debiasing Gate that dynamically arbitrates teacher-expert influence based on real-time user-item context through end-to-end meta-learning. Extensive experiments on multiple real-world datasets demonstrate that GUIDE significantly mitigates popularity bias while preserving recommendation accuracy, with state-of-the-art trade-offs among efficiency, accuracy, and fairness. The code and data are available at:https://github.29com/zx19971219/GUIDE.

IJCAI 2026

G-VTM: A Multimodal Vision-Trajectory Model for Generalized Vehicle Trajectory Prediction

Xinyue Zhang, Letian Gong, Yan Lin, Jinjun Cheng, Junlin Zhang, Guanyu Yao, Shengnan Guo, Youfang Lin, Shaojiang Wang, Huaiyu Wan

Generalized vehicle trajectory prediction across diverse junctions, including urban intersections and roundabouts, remains a fundamental task in Cooperative Vehicle–Infrastructure Systems (CVIS). This study faces two key challenges: (1) Generalize across junctions with heterogeneous map semantics and traffic behavioral patterns, where the former arises from differences in road topologies and traffic regulations, and the latter reflects diverse behavioral intentions of road users; (2) Scenario-adaptive interaction modeling, where single-modality trajectory learning captures local spatio-temporal correlation, but lacks map constraint and direction-aware interaction contexts. To overcome these challenges, we propose G-VTM, a generalized vision-trajectory model. G-VTM models fine-grained behavioral patterns and relative spatial interaction from trajectory modality. At the vision modality, G-VTM captures global map semantics while modeling scenario- and direction-aware interaction based on intuitive visual perception. Experiments on multiple real-world datasets collected by unmanned aerial vehicles (UAVs) demonstrate that our method achieves strong generalized performance under heterogeneous traffic conditions. The code is provided at https://github. com/zxyhaclyon/G-VTM.

IJCAI 2026

Dual-View Self-Supervised Pre-Training for Expert Finding

Mingqiao Zhang, Hongtao Liu, Yinghui Wang, Yumeng Wang, Qiyao Peng

Expert finding plays a crucial role in Community Question Answering platforms by routing questions to the most suitable answerers. The key challenge lies in learning high-quality representations for questions and experts. Most existing methods rely on limited supervised signals such as expert–question interactions, and typically focus on modeling only one side of the expert–question pair, suffering from data sparsity and incomplete representation. In this paper, we propose a Dual-view Self-Supervised pre-training framework for Expert Finding (SSEF) that simultaneously pre-trains expert and question representations from large-scale unlabeled data. Specifically, on the expert view, we design a self-supervised module with two data-augmentation strategies, namely historical behavior cropping and reordering, and optimize expert representations via contrastive learning over augmented sequences of historically answered questions. On the question view, we apply analogous augmentation strategies to capture intrinsic semantic differences among questions. The two view-specific modules are unified through multi-task learning with shared PLM parameters, enabling the model to capture latent semantic relatedness across views. Extensive experiments on six real-world CQA datasets demonstrate that SSEF consistently outperforms existing methods, and further analysis confirms its effectiveness under zero-shot settings and its transferability to other models.

IJCAI 2026

BotVA: Combating Social Bots via Variational Feature Augmentation and Adversarial Graph Learning

Longlong Zhang, Xi Wang, hongyi nie, Zeqing Zhang, Huixiang Zhang, Hongping Wang, Yang Liu

Social bots threaten online platforms by spreading disinformation and manipulating public discourse. Graph neural networks have emerged as effective tools for bot detection by modeling user interactions, yet two fundamental challenges limit their practical deployment: severe class imbalance where bots constitute a small minority of users, and camouflaged edges where bots forge deceptive connections to humans to evade detection. Class imbalance causes decision boundaries to shift toward the majority class, while camouflaged edges corrupt neighborhood aggregation through spurious message passing. We present BotVA, a unified framework addressing both challenges through variational feature augmentation and adversarial graph learning. Our approach makes three key contributions: (i) a conditional variational autoencoder that models minority class distributions and synthesizes semantically coherent features, effectively expanding minority support in representation space; (ii) an adversarial training paradigm where a generator simulates camouflage by injecting deceptive edges while a graph transformer discriminator with semantic attention learns to identify and downweight such perturbations; and (iii) a two-stage training strategy that pretrains the variational module before alternating generator-discriminator optimization to ensure stable convergence. Experiments on three benchmarks demonstrate that BotVA achieves state-of-the-art accuracy and F1-score, exhibits strong robustness under camouflage perturbations, and maintains competitive performance with limited supervision.

IJCAI 2026

Robust Contrastive Graph Clustering with Adaptive Local-Global Integration

Lei Zhang, Fubo Sun, Haipeng Yang, Zhong Guan, Likang Wu

Graph clustering is essential in graph analysis for revealing structural patterns and node communities. Despite recent advances in self-supervised contrastive learning that have improved clustering via structural and attribute signals, existing methods still struggle to flexibly capture high-order local structures and often overlook global semantics in complex graphs. These limitations lead to suboptimal node representations, especially in real-world graphs with fragmented structures and ambiguous cluster boundaries. To address these limitations, a contrastive graph clustering framework is proposed to jointly integrate multi-scale local structures with global semantics via attention mechanisms. At the local level, GNN-based topological signals extracted from multiple propagation depths are adaptively fused through attention-based weighting to capture multi-scale neighborhood features. At the global level, semantic prototypes derived from dynamically evolving cluster centers are adaptively aggregated through attention to guide node representations and enhance inter-cluster separability. The model is trained under a dual-view contrastive learning paradigm with a hybrid objective that combines instance-level and structure-aware losses to improve representation robustness and discrimination. Experiments on eight real-world graph datasets demonstrate that our method achieves competitive clustering performance. Code is available at https://github.com/vege12138/w2.

IJCAI 2026

Multi-Semantic Aware Self-Supervised Learning for Multi-Label Node Classification

Jiayu Zhang, Jitao Zhao, Dongxiao He, Cuiying Huo, Zhiyong Feng

Graph self-supervised learning aims to mine intrinsic signals from graph data itself to train models. It enables the acquisition of high-quality representations without manual annotations, making it suitable for various label-scarce scenarios and thus garnering substantial interest. Existing graph self-supervised methods typically assume that each node possesses a single semantic meaning. However, many real-world entities exhibit multiple semantics, where a single node may simultaneously belong to multiple categories. Since existing approaches are mainly designed for single-semantic settings, they often struggle to accurately capture the multiple semantics within nodes. Meanwhile, existing multi-label graph learning methods mainly depend on extensive manual annotations, which are costly and of limited applicability. To address this, we make a bold attempt to extend graph self-supervised learning to multi-label graphs. It is particularly challenging, as lacking labels makes it difficult to identify multiple semantics, let alone determine the number of underlying categories for nodes. To this end, we propose a Multi-semantic Aware Self-Supervised pretraining method (MASS) for multi-label graphs. Specifically, we propose a multi-pseudo-label decomposition mechanism, enabling adaptive learning of multiple semantics within nodes without annotations. Extensive results validates the effectiveness of MASS, and it achieved competitive performances with supervised baselines.

IJCAI 2026

Unsupervised Graph-Level Anomaly Detection via Multi-granular Graph Structure Learning

Ge Zhang, Huimei Li, Guohao Sun, Xiu Fang, Xixun Lin, Xiaobao Wang, Pengfei Jiao, Liang Yang

Graph-level anomaly detection (GLAD) aims to identify graphs that deviate from the majority in a dataset of graphs. Existing methods typically adopt either a global aggregation perspective that summarizes nodes within a graph into a representation vector, or a subgraph-oriented perspective which regards certain local subgraphs as indicators of anomalous properties. However, both paradigms operate from a fixed perspective and may fail to identify anomalous graphs whose discriminative characteristics manifest at multiple levels of structure granularity, where each level features the coarsened graphs at a specific granularity. In this paper, we propose M-GLAD, an unsupervised GLAD method via multi-granular graph structure learning. M-GLAD is grounded in the Prototype-guided Multi-granular Information Bottleneck (PMIB) principle. PMIB aims to measure the mutual information between coarsened graphs at the specific level of structure granularity and the learnable granularity-specific prototypes that summarize normal patterns of graphs at each granularity. The multiple levels of structure granularity of graphs and prototypes are abstracted by a graph structure abstraction module. This formulation enables granularity-aware anomaly scoring that considers anomalous characteristics across different levels of structure granularity. Extensive experiments on eight real-world graph datasets demonstrate that M-GLAD achieves superior performance over competitive baselines.

IJCAI 2026

CASE-Net: Deep Spatio-Temporal Representation Learning via Causal Attention and Channel Recalibration for Multivariate Time Series Classification

Fan Zhang, Yating Cui, Hua Wang

Multivariate time series (MTS) classification is foundational to pervasive computing and financial analysis, yet existing multi-scale paradigms are often constrained by suboptimal representation fidelity. We identify two critical bottlenecks: temporal non-causality in standard encoders that induces temporal confounding in non-stationary dynamics, and the absence of explicit channel saliency mechanisms that allows noise to contaminate the latent space. To address these challenges, we propose the Causal Attention and Spatio-temporal Encoder Network (CASE-Net), an architecture designed for structural manifold pre-conditioning. CASE-Net synergizes a Causal Temporal Encoder, which enforces physical arrow-of-time constraints via masked self-attention and causal convolutions, with an Adaptive Channel Recalibration module functioning as an information bottleneck to suppress detrimental noise. Comprehensive evaluations across six heterogeneous domains demonstrate that CASE-Net establishes new state-of-the-art benchmarks on four tasks, achieving a peak accuracy of 98.6% on the AWR dataset and superior robustness in non-stationary regimes.

IJCAI 2026

PDFlow: Popularity-Debiased Flow Matching for Sequential Recommendation

Zican Yang, Zeyu Li, Cong He, Xiaotao Wu, Guanfeng Liu, Pengpeng Zhao

Generative models have emerged as a powerful paradigm in sequential recommendation due to their superior distribution modeling. However, long-tail data distributions inevitably induce popularity bias, as iterative generation trajectories gravitate toward dense clusters of popular items. Current debiasing methods struggle to enforce consistent step-wise constraints, rendering them ineffective for the multi-step process of generative recommendation. To address this challenge, we propose a method called Popularity-Debiased Flow Matching for sequential recommendation (PDFlow). Specifically, grounded in the flow matching framework, PDFlow integrates a residual popularity vector field beyond the main flow to model the popularity debiasing flow at each step. This enables real-time trajectory intervention to adjust the generative path, preventing the model from overemphasizing popular items. Subsequently, we incorporate cross-popularity alignment loss, using co-occurrence patterns within the same sequence to align features of both popular and tail items, thereby mitigating popularity-driven distribution separation. Jointly, trajectory correction prevents popularity overfitting while cross-popularity alignment ensures latent fairness, effectively mitigating popularity bias. Extensive experiments on four real-world datasets demonstrate the effectiveness of PDFlow in improving long-tail coverage and overall recommendation performance. All codes and datasets are available at https://github.com/eqmll/PDFlow.

IJCAI 2026

DGCPath: Distribution-Aware Generative Contrastive Framework for Self-supervised Path Representation Learning

Sean Bin Yang, Hao Miao, Zongyi Xu, Jilin Hu, Xiangmeng Wang, Hua Lu, Bin Yang, Christian S. Jensen

Due to proliferation of vehicle trajectory data from advanced sensing technologies, path representation learning has become a pivotal task in intelligent transportation systems. Although existing self-supervised approaches work to some extent, their dependence on deterministic contrastive learning paradigms and handcrafted view augmentation strategies inherently restricts cross-scenario generalization capabilities. To address these limitations, we present DGCPath – an innovative Distribution-aware Generative Contrastive learning framework for Path representation. This architecture establishes a synergistic connection between generative modeling and distributional contrastive learning, enabling the acquisition of robust and transferable feature embeddings. Specifically, our framework incorporates: (1) a diffusion-based view generator that autonomously produces semantically coherent yet diversified trajectory views from Gaussian noise distributions; (2) a variational contrastive mechanism enforcing latent feature alignment at the distribution level, transcending conventional instance-wise consistency; and (3) a novel generative cross-supervision module that reinforces view-level consistency through cross-view reconstruction learning. Comprehensive evaluations on three real-world trajectory datasets demonstrate DGCPath's superior performance over state-of-the-art baselines in two distinct downstream tasks, validating its enhanced generalization capacity and representation effectiveness.

IJCAI 2026

A Unified Prompt for Enhancing Heterogeneous Graph Pre-training via Edge-based Message Passing

Fengyu Yan, Xiaobao Wang, Qianhua Tang, Dongxiao He, Di Jin

Inspired by natural language processing prompt learning, recent heterogeneous graph prompt-tuning methods have been developed to better align pre-trained models with downstream tasks. However, existing heterogeneous prompt methods primarily focus on holistic framework design, causing prompts to heavily depend on specific pre-trained models and thus limiting generalization. To address this, we focus on the common encoder module shared across pre-trained models and the multi-relational edge structure unique to heterogeneous graphs, proposing a novel heterogeneous graph prompt-tuning method named HGMRP. It simultaneously resolves the issue of inconsistent objectives during prompt optimization when messages propagate across different relational edges. By assigning learnable prompt vectors to different types of edges, HGMRP integrates multi-relational prompts directly into the message-passing process. Furthermore, we extend single-relation prompts into a composite structure that includes both relation-specific and shared components, thereby enhancing expressive capability. Extensive experiments conducted on four widely used heterogeneous graph datasets show that HGMRP can adapt to various types of pre-trained models and significantly outperforms existing heterogeneous prompt methods, validating its effectiveness and superiority.

IJCAI 2026

scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing

Ping Xu, Pengjiang Li, Tian Du, Zaitian Wang, Jiawei Gu, Zhiyuan Ning, Ziyue Qiao, Pengfei Wang, Yuanchun Zhou

Clustering is fundamental to scRNA-seq analysis, serving as a cornerstone for identifying cell populations and resolving tissue heterogeneity. However, existing methods focus on mining numerical statistical patterns, suffering from semantic agnosticism by neglecting the intrinsic biological functions encoded by genes. While Large Language Models (LLMs) offer promising semantic capabilities, their direct adaptation to cell clustering is hindered by the structural mismatch between generative pre-training objectives and discriminative downstream tasks. To bridge this gap, we propose scLLM-DSC, a novel LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering framework. Diverging from data-driven paradigms, scLLM-DSC establishes a semantically-grounded representation by synergizing two views: a Knowledge-Driven Semantic View derived from NCBI gene priors and contextualized Cell2Sentence embeddings, and a Structure-Aware Topological View extracted via a graph-guided encoder. Crucially, we introduce a cross-modal contrastive alignment mechanism to enforce consistency between biological semantics and transcriptomic features within a unified latent space. Extensive benchmarks demonstrate that scLLM-DSC significantly outperforms eleven state-of-the-art baselines in clustering accuracy.

IJCAI 2026

Snap2Review: Vision-Grounded Retrieval and Pairwise Preference Alignment for Personalized Reviews Generation

Honggyan Xu, Lei Zou, Dexiang Zhao, Xin Fan, Chuanwen Luo

Personalized review generation is vital for e-commerce engagement, yet integrating user-provided images while maintaining distinct user personas remains a critical challenge. Existing methods often struggle to bridge the semantic gap between objective visual signals and subjective linguistic patterns, frequently resulting in hallucinations or homogenized, impersonal content. To address these limitations, we propose Snap2Review, a novel framework that harmonizes cross-modal retrieval with fine-grained preference learning. First, we optimize a cross-modal retriever that maps visual features directly to semantically related historical reviews, using these interactions as precise anchors to strictly ground the generation. Second, we introduce Tri-DPO, which employs a stratified negative sampling strategy with varying difficulty levels, ranging from obvious errors to subtle stylistic mismatches, to force the model to discern fine-grained user preferences. By distinguishing these fine-grained nuances, our model moves beyond generic praise to accurately mimic the user's authentic writing style. Extensive experiments on real-world datasets demonstrate that Snap2Review significantly outperforms strong baselines in both relevance and personalization metrics.

IJCAI 2026

Three Minds, One Student: Online Multi-Teacher Knowledge Distillation for Multimodal Recommenders

Hangtong Xu, Yuanbo Xu, En Wang

Existing multimodal recommendation models using complex fusion mechanisms (e.g., attention) or multi-stage processes (e.g., early or late fusion) integrate different modalities. However, attention-based adaptive fusion is prone to shortcut learning, where dominant collaborative signals (ID) can overshadow other modalities. This dominance affects the entire fusion process: early fusion often amplifies biases driven by identity, while late fusion struggles to extract preference-relevant signals from misaligned modalities, even with alignment regularization. To address these issues, we propose Multi-Teacher Single-Student Online Distillation for Multimodal Recommendation (MTS2-4MM), which reframes the multimodal recommendation task from direct fusion to controllable knowledge transfer. Specifically, we construct multiple teachers to specialize in complementary perspectives, and a unified student distills their guidance via objectives at both the ranking and representation levels. This design explicitly controls modality contributions, improves robustness to modality noise and misalignment. Furthermore, we design a Modality-specific Preference Extractor to explicitly extract user preferences across different modalities equally. Extensive experiments across five real-world datasets demonstrate that MTS2-4MM consistently outperforms state-of-the-art baselines, achieving improvements of up to 7.22%.

IJCAI 2026

Band Together: Untargeted Adversarial Training with Multimodal Coordination Against Evasion-Based Promotion Attacks

Guanmeng Xian, Ning Yang, Philip S. Yu

Multimodal recommender systems exploit visual and textual signals to alleviate data sparsity, but this also makes them more vulnerable to evasion-based promotion attacks. Existing defenses are largely limited to single-modal settings and mainly focus on poisoning-based threats, leaving evasion-based threats underexplored. In this work, we first identify a cross-modal gradient mismatch under the multi-user promotion setting, where visual and textual perturbations are optimized in inconsistent directions due to the dominance of distinct user groups. This phenomenon dilutes the attack effectiveness and leads robust training to underestimate worst-case risks. To address this issue, we propose Untargeted Adversarial Training with Multimodal Coordination (UAT-MC). UAT-MC tackles the challenge of unknown targeted items in evasion-based attacks (as opposed to poisoning-based attacks) by treating all items as potential targets, and introduces a gradient alignment mechanism to explicitly correct this mismatch. This design ensures synchronized perturbations across modalities, thereby maximizing adversarial strength for robust training. Extensive experiments demonstrate that UAT-MC significantly improves robustness against promotion attacks while maintaining acceptable recommendation performance under the defense–accuracy trade-off. Code is available at https://github.com/ gmXian/UAT-MC.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.