对2025年的 推荐系统论文进行一波收集,给各位初学者和算法大佬作为灵感来源,后续专栏会继续更新论文解读,根据评论不断补充,欢迎大家三连~
WSDM 2025
Review-Based Hyperbolic Cross-Domain Recommendation(基于评论的双曲跨域推荐)
Beyond Message-Passing: Generalization of Graph Neural Networks via Feature Perturbation(超越消息传递:通过特征扰动对图神经网络进行泛化)
LOGIN: A Large Language Model Consulted Graph Neural Network Training Framework(LOGIN:一个基于大型语言模型的图神经网络训练框架)
Dynamic Graph Transformer with Correlated Spatial-Temporal Positional Encoding(具有相关时空位置编码的动态图变换器)
S-Diff: An Anisotropic Diffusion Model for Collaborative Filtering in Spectral Domain(S-Diff:一种频谱域协同过滤的各向异性扩散模型)
Maintaining k-MinHash Signatures over Fully-Dynamic Data Streams with Recovery(在全动态数据流上维护带恢复的k-MinHash签名)
Hyperdimensional Representation Learning for Node Classification and Link Prediction(用于节点分类和链路预测的高维表示学习)
FedGF: Enhancing Structural Knowledge via Graph Factorization for Federated Graph Learning(FedGF:通过图分解增强联邦图学习的结构知识)
Prospective Multi-Graph Cohesion for Multivariate Time Series Anomaly Detection(用于多变量时间序列异常检测的前瞻性多图凝聚)
Incomplete Multi-view Clustering via Local Reasoning and Correlation Analysis(通过局部推理和相关性分析的不完备多视图聚类)
Large Language Model driven Policy Exploration for Recommender Systems(用于推荐系统的大型语言模型驱动策略探索)
The Initial Screening Order Problem(初始筛选顺序问题)
Cross-Domain Pre-training with Language Models for Transferable Time Series Representations(利用语言模型进行跨域预训练以获得可迁移时间序列表示)
Empowering Time Series Classification with Multimodal Language Modeling(利用多模态语言建模增强时间序列分类)
Unsupervised Robust Cross-Lingual Entity Alignment via Neighbor Triple Matching with Entity and Relation Texts(通过实体和关系文本的邻域三元组匹配进行无监督鲁棒跨语言实体对齐)
MedTransTab: Advancing Medical Cross-Table Tabular Data Generation(MedTransTab:推进医学跨表表格数据生成)
Q-DISCO: Query-Centric Densest Subgraphs in Networks with Opinion Information(Q-DISCO:带有意见信息的网络中以查询为中心的最密子图)
Gradient Deconfliction via Orthogonal Projections onto Subspaces For Multi-task Learning(通过子空间正交投影进行多任务学习的梯度去冲突)
Self-supervised Time-aware Heterogeneous Hypergraph Learning for Dynamic Graph-level Classification(用于动态图级别分类的自监督时间感知异构超图学习)
Combating Heterogeneous Model Biases in Recommendations via Boosting(通过提升来对抗推荐中的异构模型偏差)
Robustness Verification of Deep Graph Neural Networks Tightened by Linear Approximation(通过线性逼近强化的深度图神经网络鲁棒性验证)
Teach Me How to Denoise: a Universal Framework for Denoising Multi-modal Recommender Systems via Guided Calibration(教我如何去噪:一个通过引导校准对多模态推荐系统去噪的通用框架)
Dynamic Interaction-Driven Intent Evolver with Semantic Probability Distributions(具有语义概率分布的动态交互驱动意图演化器)
DDualSE: Decoupled Dual-head Squeeze and Excitation Attention for Sequential Recommendation(DDualSE:用于序列推荐的解耦双头挤压与激励注意力机制)
Do Stubborn Users Always Cause More Polarization and Disagreement? A Mathematical Study(顽固用户总是会导致更多极化和分歧吗?一项数学研究)
HHGT: Hierarchical Heterogeneous Graph Transformer for Heterogeneous Graph Representation Learning(HHGT:用于异构图表示学习的层次异构图变换器)
A Sublinear Algorithm for Approximate Shortest Paths in Large Networks(大型网络中近似最短路径的次线性算法)
Privacy-Preserving Orthogonal Aggregation for Guaranteeing Gender Fairness in Federated Recommendation(联邦推荐中保障性别公平的隐私保护正交聚合)
Writing Style Matters: An Examination of Bias and Fairness in Information Retrieval Systems(写作风格很重要:信息检索系统中偏差与公平性的考察)
CIMAGE: Exploiting the Conditional Independence in Masked Graph Auto-encoders(CIMAGE:利用掩码图自动编码器中的条件独立性)
RSM: Reinforced Subgraph Matching Framework with Fine-grained Operation based Search Plan(RSM:基于细粒度操作搜索计划的强化子图匹配框架)
BAKER: Bayesian Kernel Uncertainty in Domain-Specific Document Modelling(BAKER:特定领域文档建模中的贝叶斯核不确定性)
Edge Classification on Graphs: New Directions in Topological Imbalance(图上的边分类:拓扑不平衡的新方向)
Bridging Source and Target Domains via Link Prediction for Unsupervised Domain Adaptation on Graphs(通过链路预测连接源域和目标域以实现图上的无监督域适应)
Hawkes Point Process-Enhanced Dynamic Graph Neural Networks(霍克斯点过程增强的动态图神经网络)
Optimizing Blockchain Analysis: Tackling Temporality and Scalability with an Incremental Approach with Metropolis-Hastings Random Walks(优化区块链分析:通过基于 metropolis - hastings 随机游走的增量方法处理时间性和可扩展性)
SCONE: A Novel Stochastic Sampling to Generate Contrastive Views and Hard Negative Samples for Recommendation(SCONE:一种用于生成推荐对比视图和硬负样本的新型随机采样方法)
Towards Personalized Federated Multi-Scenario Multi-Task Recommendation(迈向个性化联邦多场景多任务推荐)
Revisiting Fake News Detection: Towards Temporality-aware Evaluation by Leveraging Engagement Earliness(重新审视假新闻检测:通过利用早期参与实现时间感知评估)
Graph Disentangle Causal Model: Enhancing Causal Inference in Networked Observational Data(图解耦因果模型:增强网络观测数据中的因果推理)
ESA: Example Sieve Approach for Multi-Positive and Unlabeled Learning(ESA:用于多正例和无标签学习的示例筛选方法)
Context Embeddings for Efficient Answer Generation in RAG(用于检索增强生成中高效答案生成的上下文嵌入)
An aspect performance-aware hypergraph neural network for review-based recommendation(一种用于基于评论推荐的方面性能感知超图神经网络)
Balancing Revenue and Privacy with Signaling Schemes in Online Ad Auctions(通过在线广告拍卖中的信号机制平衡收益和隐私)
A Contrastive Framework with User, Item and Review Alignment for Recommendation(一种具有用户、项目和评论对齐的推荐对比框架)
Sequential diversification with provable guarantees(具有可证明保证的序列多样化)
DeMBR: Denoising Model with Memory Pruning and Semantic Guidance for Multi-Behavior Recommendation(DeMBR:用于多行为推荐的具有记忆修剪和语义引导的去噪模型)
Progressive Tasks Guided Multi-Source Network for Customer Lifetime Value Prediction in Online Advertising(用于在线广告中客户终身价值预测的渐进任务引导多源网络)
Mining Topics towards ChatGPT using Disentangled Contextualized-neural Topic Model(利用解耦的上下文神经主题模型挖掘面向ChatGPT的主题)
Exploration and Exploitation of Hard Negative Samples for Cross-Domain Sequential Recommendation(跨域序列推荐中硬负样本的探索与利用)
Simple Graph Neural Networks for Recommendation(用于推荐的简单图神经网络)
Enhancing Code Search Intent with Programming Context Exploration(通过编程上下文探索增强代码搜索意图)
Neo-TKGC: Enhancing Temporal Knowledge Graph Completion with Integrated Node Weights and Future Information(Neo-TKGC:通过集成节点权重和未来信息增强时态知识图补全)
AMLCDR: An Adaptive Meta-Learning Model for Cross-Domain Recommendation by Aligning Preference Distributions(AMLCDR:一种通过对齐偏好分布进行跨域推荐的自适应元学习模型)
HACD: Harnessing Attribute Semantics and Mesoscopic Structure for Community Detection(HACD:利用属性语义和介观结构进行社区检测)
GAMED: Knowledge Adaptive Multi-Experts Decoupling for Multimodal Fake News Detection(GAMED:用于多模态假新闻检测的知识自适应多专家解耦)
Oracle-guided Dynamic User Preference Modeling for Sequential Recommendation(用于序列推荐的基于神谕引导的动态用户偏好建模)
IMPO: Interpretable Memory-based Prototypical Pooling(IMPO:可解释的基于记忆的原型池化)
Improving Scientific Document Retrieval with Concept Coverage-based Query Set Generation(通过基于概念覆盖的查询集生成改进科学文献检索)
Personalised Outfit Recommendation via History-aware Transformers(通过历史感知变换器进行个性化服装推荐)
LLM Simulator for Online Billion-Scale Item Cold-Start Recommendation(用于在线十亿规模项目冷启动推荐的大型语言模型模拟器)
DTPN: A Diffusion-based Traffic Purification Network for Tor Website Fingerprinting(DTPN:一种用于Tor网站指纹识别的基于扩散的流量净化网络)
Density-aware and Cluster-based Federated Anomaly Detection on Data Streams(数据流上基于密度感知和聚类的联邦异常检测)
Predicting Eviction Status Using Airbnb Data in the Absence of Ground-Truth Eviction Records(在缺乏真实驱逐记录的情况下利用爱彼迎数据预测驱逐状态)
Training MLPs on Graphs without Supervision(无监督地在图上训练多层感知机)
Explainable CTR prediction via LLM reasoning(通过大型语言模型推理进行可解释的点击率预测)
Facet-Aware Multi-Head Mixture-of-Experts Model for Sequential Recommendation(用于序列推荐的方面感知多头混合专家模型)
Adaptive Graph Enhancement for Imbalanced Multi-relation Graph Learning(用于不平衡多关系图学习的自适应图增强)
Temporal Linear Item-Item Model for Sequential Recommendation(用于序列推荐的时间线性项目 - 项目模型)
DimeRec: A Unified Framework for Enhanced Sequential Recommendation via Generative Diffusion Models(DimeRec:一种通过生成扩散模型增强序列推荐的统一框架)
An Edge-Based Decomposition Framework for Temporal Networks(用于时间网络的基于边的分解框架)
Fusion Matters: Learning Fusion in Deep Click-through Rate Prediction Models(融合很重要:深度学习点击率预测模型中的融合学习)
VARIUM: Variational Autoencoder for Multi-Interest Representation with Inter-User Memory(VARIUM:具有用户间记忆的多兴趣表示变分自编码器)
Graph Size-imbalanced Learning with Energy-guided Structural Smoothing(具有能量引导结构平滑的图尺寸不平衡学习)
Polaris: Sampling from the Multigraph Configuration Model with Prescribed Color Assortativity(Polaris:从具有规定颜色同配性的多重图配置模型中采样)
Spectrum-based Modality Representation Fusion Graph Convolutional Network for Multi-modal Recommendation(用于多模态推荐的基于频谱的模态表示融合图卷积网络)
Towards Reliable Latent Knowledge Estimation in LLMs: Zero-Prompt Many-Shot Based Factual Knowledge Extraction(迈向大型语言模型中可靠的潜在知识估计:基于零提示多示例的事实知识提取)
Improving FIM Code Completions via Context & Curriculum Based Learning(通过基于上下文和课程的学习改进函数式接口管理代码补全)
Sequentially Diversified and Accurate Recommendations in Chronological Order for a Series of Users(针对一系列用户按时间顺序进行顺序多样化且准确的推荐)
Adjacent Neighborhood Transformer-based Diffusion Model for Anomaly Detection under Incomplete Industrial Data Sources(基于邻域变换器的不完全工业数据源下异常检测扩散模型)
Untapping the Power of Indirect Relationships in Entity Summarization(挖掘实体摘要中间接关系的力量)
Improving CTR Prediction with Graph-Enhanced Interest Networks for Sparse Behavior Sequences(通过图增强兴趣网络改进稀疏行为序列的点击率预测)
Heterophilic Graph Neural Networks Optimization with Causal Message-passing(通过因果消息传递优化异质图神经网络)
HaGAR: Hardness-aware Generative Adversarial Recommender(HaGAR:硬度感知生成对抗推荐器)
UIPN: User Intent Profiling Network for Multi Behavior Modeling in CTR Prediction(UIPN:用于点击率预测中多行为建模的用户意图剖析网络)
MoKGNN: Boosting Graph Neural Networks via Mixture of Generic and Task-Specific Language Models(MoKGNN:通过通用和特定任务语言模型的混合提升图神经网络)
DLCRec: A Novel Approach for Managing Diversity in LLM-Based Recommender Systems(DLCRec:一种用于管理基于大型语言模型推荐系统多样性的新方法)
Reindex-Then-Adapt: Improving Large Language Models for Conversational Recommendation(重新索引然后适应:改进用于对话推荐的大型语言模型)
Adaptive Loss-based Curricula for Neural Team Recommendation(用于神经团队推荐的基于自适应损失的课程)
Demystify Epidemic Containment in Directed Networks: Theory and Algorithms(揭开有向网络中疫情遏制的神秘面纱:理论与算法)
Does Memorization Affect LLMs’ Social Reasoning? An Evaluation on Seen and Unseen Queries(记忆是否影响大型语言模型的社会推理?对已见和未见查询的评估)
Beyond Answers: Transferring Reasoning Capabilities to Smaller LLMs Using Multi-Teacher Knowledge Distillation(超越答案:利用多教师知识蒸馏将推理能力转移到较小的大型语言模型)
Inductive Graph Few-shot Class Incremental Learning(归纳图小样本类别增量学习)
How Do Recommendation Models Amplify Popularity Bias? An Analysis from the Spectral Perspective(推荐模型如何放大流行度偏差?从频谱角度的分析)
Lighter And Better: Flexible Context Adaptation For Retrieval Augmented Generation(更轻更好:用于检索增强生成的灵活上下文适应)
RetriEVAL: Evaluating Text Generation with Contextualized Lexical Match(RetriEVAL:通过上下文词汇匹配评估文本生成)
Effective Learning in the Presence of Unstructured Interactions(存在非结构化交互时的有效学习)
Generating Diverse Criteria On-the-Fly to Improve Point-wise LLM Rankers(即时生成多样化标准以改进逐点大型语言模型排名器)
Quam: Adaptive Retrieval through Query Affinity Modelling(Quam:通过查询亲和性建模进行自适应检索)
ProCC: Programmatic Reinforcement Learning for Efficient and Transparent TCP Congestion Control(ProCC:用于高效透明TCP拥塞控制的程序化强化学习)
UniGLM: Training One Unified Language Model for Text-Attributed Graphs Embedding(UniGLM:训练一个用于文本属性图嵌入的统一语言模型)
Your causal self-attentive recommender hosts a lonely neighborhood(你的因果自注意力推荐器存在孤立邻域)
HTEA: Heterogeneity-aware Temporal Entity Alignment(HTEA:异构感知时间实体对齐)
D
2
^2
2: Customizing Two-Stage Graph Neural Networks for Early Rumor Detection through Cascade Diffusion Prediction(D²:通过级联扩散预测定制用于早期谣言检测的两阶段图神经网络)
Heterogeneous Graph Diffusion Model(异构图扩散模型)
Heterogeneous Disentangled Collaborative Filtering(异构解耦协同过滤)