论文检索

Jinfeng Xu 0003, Zheyu Chen 0003, Jinze Li 0001, Shuo Yang 0011, Hewei Wang 0001, Yijie Li 0003, Mengran Li 0001, Puzhen Wu, Edith C. H. Ngai

The data sparsity problem significantly hinders the performance of recommender systems, as traditional models rely on limited historical interactions to learn user preferences and item properties. While incorporating multimodal information can explicitly represent these preferences and properties, existing works often use it only as side information, failing to fully leverage its potential. In this paper, we propose MDVT, a model-agnostic approach that constructs multimodal-driven virtual triplets to provide valuable supervision signals, effectively mitigating the data sparsity problem in multimodal recommendation systems. To ensure high-quality virtual triplets, we introduce three tailored warm-up threshold strategies: static, dynamic, and hybrid. The static warm-up threshold strategy exhaustively searches for the optimal number of warm-up epochs but is time-consuming and computationally intensive. The dynamic warm-up threshold strategy adjusts the warm-up period based on loss trends, improving efficiency but potentially missing optimal performance. The hybrid strategy combines both, using the dynamic strategy to find the approximate optimal number of warm-up epochs and then refining it with the static strategy in a narrow hyper-parameter space. Once the warm-up threshold is satisfied, the virtual triplets are used for joint model optimization by our enhanced pair-wise loss function without causing significant gradient skew. Extensive experiments on multiple real-world datasets demonstrate that integrating MDVT into advanced multimodal recommendation models effectively alleviates the data sparsity problem and improves recommendation performance, particularly in sparse data scenarios.

Bihan Xu, Shiwei Zhao, Runze Wu 0001, Zhenya Huang, Jiawei Wang, Zhipeng Hu, Kai Wang 0064, Haoyu Liu 0002, Tangjie Lv, Le Li 等

Within the domain of Massively Multiplayer Online (MMO) economy research, Agent-Based Modeling (ABM) has emerged as a robust tool for analyzing game economics, evolving from rule-based agents to decision-making agents enhanced by reinforcement learning. Nevertheless, existing works encounter significant challenges when attempting to emulate human-like economic activities among agents, particularly regarding agent reliability, sociability, and interpretability.In this study, we take a preliminary step in introducing a novel approach using Large Language Models (LLMs) in MMO economy simulation. Leveraging LLMs' role-playing proficiency, generative capacity, and reasoning aptitude, we design LLM-driven agents with human-like decision-making and adaptability. These agents are equipped with the abilities of role-playing, perception, memory, and reasoning, addressing the aforementioned challenges effectively. Simulation experiments focusing on in-game economic activities demonstrate that LLM-empowered agents can promote emergent phenomena like role specialization and price fluctuations in line with market rules.

Hao Xiong 0003, Yebin Yang, Huaijin Wu, Xiaoqiu Zhong, Yehui Tang 0002, Zhuo Xia, Xiaoxing Wang, Junchi Yan

We rethink the basic operations, i.e., inner product and matrix multiplication used in neural networks. A quantum-inspired alternative is proposed, utilizing the power of high-dimensional Hilbert space by devising a high-order form of tensor product. We re-parameterize the original (low-order) vectors/matrices into an expressive high-order form, without incurring extra model parameters, and the extra computational overhead is negligible (e.g., about 2%). As an in-place transparent atomic operation, we show its use in the key components in Transformers: token embeddings, attentions (query, key, value) and the MLP. Due to its inherent compatibility to vanilla multiplicative operations, we propose C2Q-SFT, i.e., classic-to-quantum (C2Q) protocol for supervised fine-tuning (SFT): it continues to train a given model by transparently replacing the standard operations with ours. As shown by our experiments, it shows advantages for both training from scratch and fine-tuning on downstream tasks across scales of LLMs. C2Q-SFT consistently outperforms standard SFT, with relative improvements on MMLU (+0.56%) and GSM8k (+0.61%). It sheds light on the innovation of operations in networks, orthogonal to the efforts on new architecture, position encoding, and training algorithms, etc. See project page at: https://github.com/Thinklab-SJTU/LLM/QI-LLM.

Amy Xin, Jinxin Liu 0002, Zijun Yao 0002, Zhicheng Lee, Shulin Cao, Lei Hou 0001, Juanzi Li

Despite the outstanding capabilities of large language models (LLMs), knowledge-intensive reasoning still remains a challenging task due to LLMs' limitations in compositional reasoning and the hallucination problem. A prevalent solution is to employ chain-of-thought (CoT) with retrieval-augmented generation (RAG), which first formulates a reasoning plan by decomposing complex questions into simpler sub-questions, and then applies iterative RAG at each sub-question. However, prior works exhibit two crucial problems: inadequate reasoning planning and poor incorporation of heterogeneous knowledge. In this paper, we introduce AtomR, a framework for LLMs to conduct accurate heterogeneous knowledge reasoning at the atomic level. Inspired by how knowledge graph query languages model compositional reasoning through combining predefined operations, we propose three atomic knowledge operators, a unified set of operators for LLMs to retrieve and manipulate knowledge from heterogeneous sources. First, in the reasoning planning stage, AtomR decomposes a complex question into a reasoning tree where each leaf node corresponds to an atomic knowledge operator, achieving question decomposition that is highly fine-grained and orthogonal. Subsequently, in the reasoning execution stage, AtomR executes each atomic knowledge operator, which flexibly selects, retrieves, and operates atomic level knowledge from heterogeneous sources. We also introduce BlendQA, a challenging benchmark specially tailored for heterogeneous knowledge reasoning. Experiments on three single-source and two multi-source datasets show that AtomR outperforms state-of-the-art baselines by a large margin, with absolute F1 score improvements of 9.4% on 2WikiMultihop and 9.5% on BlendQA. We release our code and data https://github.com/THU-KEG/AtomR.git.

Wenjia Xie, Hao Wang 0076, Minghao Fang, Ruize Yu 0001, Wei Guo 0006, Yong Liu 0020, Defu Lian, Enhong Chen

Sequential recommendation (SR), as an important branch of recommendation systems, has garnered significant attention due to its substantial commercial value. This has inspired some researchers to draw from the successful experiences of large language models to develop scaling laws for SR. However, the improvements brought by parameter expansion often reach a limit when the data scale is fixed. We have observed that existing deep learning sequence methods are typically seen as learning a unified pattern of user interactions, as they apply the same model for inference across different users, which often leads to the neglect of individual user behavior patterns. To address this, we propose conducting an independent analysis of each user's interaction sequence in SR. We initially developed the PCRec-simple, which uses KL divergence to perform a one-time optimization on each sequence after training, demonstrating that optimizing individual sequences can provide additional insights and overcome the performance bottleneck after scaling laws. Subsequently, we introduce PCRec, a sequential recommendation model that integrates real-time inference of hidden states into the model. It applies KL divergence optimization during the forward process, allowing for end-to-end optimization and addressing issues of robustness, parallelism, and optimization stability. Extensive experiments on real-world datasets show that PCRec significantly outperforms the current state-of-the-art methods. The code can be found at https://github.com/USTC-StarTeam/PCRec.

Lindong Xie, Yang Zhang, Zhixian Tang, Edward Chung 0001, Genghui Li, Zhenkun Wang 0001

Surrogate-assisted evolutionary algorithms (SAEAs) are well-suited for optimizing computationally expensive black-box problems in diverse real-world scenarios. The sample efficiency of SAEAs depends largely on the configuration of the surrogate model and sampling criteria. However, configuring these core components requires substantial manual effort and expert knowledge, limiting the broader applicability of SAEAs. To address these challenges, we propose CoE-SAEA, a novel paradigm that co-evolves large language models (LLMs) and configuration strategies to enhance SAEAs. Specifically, the paradigm consists of three populations with distinct roles: one evolves LLM prompts to generate robust configuration strategy instructions, another optimizes the configuration strategies, and the third solves the optimization problem using the selected algorithm configuration. Additionally, an exploration-exploitation module is incorporated to decide whether to explore new configuration strategies via LLMs or exploit existing ones. We empirically validate the efficacy of CoE-SAEA by comparing it to state-of-the-art algorithms across various benchmark problems and a real-world traffic signal optimization task. The source code of the proposed CoE-SAEA is publicly available at: https://github.com/ForrestXie9/CoE-SAEA.

Huan Xie, Qing Liu 0008, Chengyang Luo 0002, Yuhan Zhou 0001, Yunjun Gao

In this paper, we investigate a new problem of truss-based why-not community search. Given a k-truss community C in a graph G and a why-not vertex w ∉ C , the goal is to insert the minimum number of new edges into G to ensure that w becomes part of the k-truss community. This problem has a wide range of applications, such as friends recommendation and transportation planning. We prove that the truss-based why-not community search problem is NP-hard and propose two efficient heuristic algorithms: the expansion-based algorithm and the simulation-based algorithm. Specifically, the expansion-based algorithm incrementally inserts the edges one by one, guided by a carefully designed edge goodness function that quantifies edge quality to ensure optimal selection. In contrast, the simulation-based algorithm firstly inserts a sufficient number of edges into G to immediately include w in a k-truss community and then removes redundant edges to minimize insertions. Furthermore, we implement a set of optimizations to further enhance the efficiency of both algorithms. Extensive experiments on real-world graphs demonstrate the efficiency and effectiveness of our proposed methods and optimizations.

Bin Xiang, Bogdan Cautis, Xiaokui Xiao, Laks V. S. Lakshmanan

In this paper, we investigate the diffusion containment problem through a novel hyperparametric influence minimization model. This model integrates a hyperparametric diffusion framework into the classical influence minimization paradigm, enabling practical, flexible, and fine-grained control over diffusion dynamics via feature interventions on nodes. The objective is to minimize the diffusion from initial seeds, by optimizing the interventions on node feature values. We analyze the challenges and intrinsic properties of hyperparametric influence minimization and derive an upper-bound on the spread, which quantifies the total uncertainty of nodes remaining inactive during the diffusion process. We prove that it exhibits supermodularity in the context of the node selection problem. Based on that, we further design greedy-based algorithms to solve the problem, which outperform the state-of-the-art methods.

Hongbo Xia, Kaiqiang Yu, Shengxin Liu, Cheng Long 0001, Xun Zhou 0001

Cohesive subgraph mining is a fundamental problem in graph theory with numerous real-world applications, such as social network analysis and protein-protein interaction modeling. Among various cohesive subgraphs, the γ-quasi-clique is widely studied for its flexibility in requiring each vertex to connect to at least a γ proportion of other vertices in the subgraph. However, solving the maximum γ-quasi-clique problem is NP-hard and further complicated by the lack of the hereditary property, which makes designing efficient pruning strategies challenging. Existing algorithms, such as DDA and FastQC, either struggle with scalability or exhibit significant performance declines for small values of γ. In this paper, we propose a novel algorithm, IterQC, which reformulates the maximum γ-quasi-clique problem as a series of k-plex problems that possess the hereditary property. IterQC introduces a non-trivial iterative framework and incorporates two key optimization techniques: (1) the pseudo lower bound (pseudo LB) technique, which leverages information across iterations to improve the efficiency of branch-and-bound searches, and (2) the preprocessing technique that reduces problem size and unnecessary iterations. Extensive experiments demonstrate that IterQC achieves up to four orders of magnitude speedup and solves significantly more graph instances compared to state-of-the-art algorithms DDA and FastQC.

Zhangkai Wu, Xuhui Fan 0001, Jin Li 0028, Zhilin Zhao 0001, Hui Chen 0026, Longbing Cao

The recently proposed Bayesian Flow Networks (BFNs) show great potential in modeling parameter spaces via a diffusion process, offering a unified strategy for handling continuous, discrete data. However, these parameter diffusion models cannot learn high-level semantic representation from the parameter space since common encoders, which encode data into one static representation, can- not capture semantic changes in parameters. This motivates a new direction: learning semantic representations hidden in the param- eter spaces to characterize noisy data. Accordingly, we propose a representation learning framework named SepDiff which operates in the parameter space to obtain parameter-wise latent semantics that exhibit progressive structures. Specifically, SepDiff proposes a self-encoder to learn latent semantics directly from parameters, rather than from observations. The encoder is then integrated into parameter diffusion model, enabling representation learning with various formats of observations. Mutual information terms further promote the disentanglement of latent semantics and capture mean- ingful semantics simultaneously. We illustrate seven representation learning tasks in SepDiff via expanding this parameter diffusion model, and extensive quantitative experimental results demonstrate the superior effectiveness of SepDiff in learning parameter repre- sentation.

Yongxuan Wu, Yang Liu 0320, Xixun Lin, Hong Zhou, Yanan Cao 0001, Lixin Zou, Yanmin Shang, Yanbing Liu 0007

Cross-domain recommendation (CDR) has gained significant attention for its ability to address data sparsity issue. However, most existing CDR methods focus primarily on improving recommendation accuracy while largely overlooking fairness considerations, which can lead to biased outcomes and unfair treatment of different user groups. To solve this critical problem, we investigate whether fairness can be transferred from the source domain to the target domain. Our analysis suggests that fairness can be effectively transferred if the fairness of the source domain is ensured and the distributions of the source and target domains are well aligned. Based on this, we propose the FairCDR, a novel framework that can achieve the knowledge transfer of fairness and user preferences simultaneously. FairCDR owns two phases: single-domain fairness guarantee and inter-domain distribution alignment. In the first phase, we employ an adversarial learning-based recommender (ALR) to disentangle user preferences from sensitive attributes in the source domain. In the second phase, we introduce a new mutual learning-based diffusion model (MLDiff), which engages in mutual learning with ALR to progressively align the distributions of the source and target domains. This improves ALR's adaptability to distribution shifts, ultimately ensuring fairness and recommendation performance in the target domain. Extensive experiments on multiple real-world cross-domain datasets demonstrate that FairCDR surpasses existing strong baselines in both fairness and recommendation quality.

Xuansheng Wu, Wenhao Yu 0002, Xiaoming Zhai, Ninghao Liu 0001

Modern text classification methods heavily rely on contextual embeddings from large language models (LLMs). Compared to human-engineered features, these embeddings provide automatic and effective representations for classification model training. However, they also introduce a challenge: we lose the ability to manually remove unintended features, such as sensitive or task-irrelevant features, to guarantee regulatory compliance or improve the generalizability of classification models. This limitation arises because LLM embeddings are opaque and difficult to interpret. In this paper, we propose a novel framework to identify and regularize unintended features in the LLM latent space. Specifically, we first pre-train a sparse autoencoder (SAE) to extract interpretable features from LLM latent spaces. To ensure the SAE can capture task-specific features, we further fine-tune it on task-specific datasets. In training the classification model, we propose a simple and effective regularizer, by minimizing the similarity between the classifier weights and the identified unintended feature, to remove the impact of these unintended features on classification. We evaluate the proposed framework on three real-world tasks, including toxic chat detection, reward modeling, and disease diagnosis. Results show that the proposed self-regularization framework can improve the classifier's generalizability by regularizing those features that are not semantically correlated to the task. This work pioneers controllable text classification on LLM latent spaces by leveraging interpreted features to address generalizability, fairness, and privacy challenges. The code and data are publicly available at https://github.com/JacksonWuxs/Controllable_LLM_Classifier.

Xuan Wu, Yizheng Zhao

Neuro-symbolic computing aims to integrate neural learning with symbolic reasoning to address the fundamental challenge of symbol grounding. While neural networks excel at pattern recognition, they struggle to maintain logical consistency. Conversely, symbolic systems provide formal reasoning capabilities but lack mechanisms for handling perceptual uncertainty. This paper introduces EmALC , a novel neuro-symbolic framework that bridges neural perception with symbolic logic through differentiable fuzzy semantics. Our approach addresses a key limitation of existing methods: while previous neuro-symbolic approaches like Logic Tensor Networks employ first-order fuzzy logic, where key reasoning problems are undecidable, EmALC ensures decidable reasoning by leveraging a fuzzy variant of ALC -- a decidable fragment of first-order logic. Unlike previous approaches that often compromise logical soundness for learning capability, EmALC maintains provable semantic consistency through a hierarchical loss function while mitigating reasoning shortcuts via rule-based revision strategies. Experimental evaluation demonstrates EmALC's effectiveness: on ontology revision tasks, it achieves 100% success rate in correcting masked groundings while preserving semantic integrity; on semantic image interpretation tasks, it improves object classification F1-scores by up to 5.56% through ontology-guided knowledge revision.

Xuan Wu 0004, Di Wang 0004, Chunguo Wu, Lijie Wen 0001, Chunyan Miao, Yubin Xiao, You Zhou 0008

Recent studies exploited Large Language Models (LLMs) to autonomously generate heuristics for solving Combinatorial Optimization Problems (COPs), by prompting LLMs to first provide search directions and then derive heuristics accordingly. However, the absence of task-specific knowledge in prompts often leads LLMs to provide unspecific search directions, obstructing the derivation of well-performing heuristics. Moreover, evaluating the derived heuristics remains resource-intensive, especially for those semantically equivalent ones, often requiring omissible resource expenditure. To enable LLMs to provide specific search directions, we propose the Hercules algorithm, which leverages our designed Core Abstraction Prompting (CAP) method to abstract the core components from elite heuristics and incorporate them as prior knowledge in prompts. We theoretically prove the effectiveness of CAP in reducing unspecificity and provide empirical results in this work. To reduce computing resources required for evaluating the derived heuristics, we propose few-shot Performance Prediction Prompting (PPP), a first-of-its-kind method for the Heuristic Generation (HG) task. PPP leverages LLMs to predict the fitness values of newly derived heuristics by analyzing their semantic similarity to previously evaluated ones. We further develop two tailored mechanisms for PPP to enhance predictive accuracy and determine unreliable predictions, respectively. The use of PPP makes Hercules more resource-efficient and we name this variant Hercules-P. Extensive experiments across four HG tasks, five COPs, and eight LLMs demonstrate that Hercules outperforms the state-of-the-art LLM-based HG algorithms, while Hercules-P excels at minimizing required computing resources. In addition, we illustrate the effectiveness of CAP, PPP, and the other proposed mechanisms by conducting relevant ablation studies.

Wei Wu 0045, Chao Wang 0086, Liyi Chen 0001, Mingze Yin, Yiheng Zhu 0002, Kun Fu 0002, Jieping Ye, Hui Xiong 0001, Zheng Wang 0027

Proteins, as essential biomolecules, play a central role in biological processes, including metabolic reactions and DNA replication. Accurate prediction of their properties and functions is crucial in biological applications. Recent development of protein language models (pLMs) with supervised fine tuning provides a promising solution to this problem. However, the fine-tuned model is tailored for particular downstream prediction task, and achieving general-purpose protein understanding remains a challenge. In this paper, we introduce Structure-Enhanced Protein Instruction Tuning (SEPIT) framework to bridge this gap. Our approach incorporates a novel structure-aware module into pLMs to enrich their structural knowledge, and subsequently integrates these enhanced pLMs with large language models (LLMs) to advance protein understanding. In this framework, we propose a novel instruction tuning pipeline. First, we warm up the enhanced pLMs using contrastive learning and structure denoising. Then, caption-based instructions are used to establish a basic understanding of proteins. Finally, we refine this understanding by employing a mixture of experts (MoEs) to capture more complex properties and functional information with the same number of activated parameters. Moreover, we construct the largest and most comprehensive protein instruction dataset to date, which allows us to train and evaluate the general-purpose protein understanding model. Extensive experiments on both open-ended generation and closed-set answer tasks demonstrate the superior performance of SEPIT over both closed-source general LLMs and open-source LLMs trained with protein knowledge.

Kun Wu 0011, Wendy Hui Wang

Graph unlearning (GU) enables data owners to remove specific data from a trained Graph Neural Network (GNN). However, a dishonest model provider may cheat on the unlearning process. This paper focuses on a specific type of cheating behavior in GU, namely incomplete edge unlearning, where the model provider removes only a subset of the requested edges from the trained GNN. We introduce PANDA, the first probabilistic GU verification framework, to detect such cheating behaviors. PANDA identifies a set of nodes, called token nodes, in the graph, and injects a set of fake edges, referred to as challenge edges, into the model to manipulate the predictions of token nodes. A key property of the challenge edges is that removing only a subset of them from the poisoned model does not change the token nodes' prediction. Then by requesting the removal of the challenge edges and observing the change in the token nodes' predictions, the verifier can assess the likelihood that the model provider has engaged in incomplete unlearning. To develop PANDA, we design a novel algorithm to identify the token nodes and generate their associated challenge edges. We rigorously quantify the verification probabilities achieved by PANDA. Our extensive empirical studies demonstrate the efficiency and effectiveness of PANDA in detecting incomplete edge unlearning across a variety of GNN models and unlearning algorithms. Furthermore, we show that PANDA exhibits strong robustness against state-of-the-art detection methods for graph adversarial perturbations. Our code and datasets are available at https://github.com/kunwu522/unlearning-verification-gnn.

Fang Wu 0002, Zhengyuan Zhou, Shuting Jin, Xiangxiang Zeng, Jure Leskovec, Jinbo Xu

Therapeutic peptides show promise in targeting previously undruggable binding sites, with recent advancements in deep generative models enabling full-atom peptide co-design for specific protein receptors. However, the critical role of molecular surfaces in protein-protein interactions (PPIs) has been underexplored. To bridge this gap, we propose an omni-design peptides generation paradigm, called SurfFlow, a novel surface-based generative algorithm that enables comprehensive co-design of sequence, structure, and surface for peptides. SurfFlow employs a multi-modality conditional flow matching (CFM) architecture to learn distributions of surface geometries and biochemical properties, enhancing peptide binding accuracy. Evaluated on the comprehensive PepMerge benchmark, SurfFlow consistently outperforms full-atom baselines across all metrics. These results highlight the advantages of considering molecular surfaces in de novo peptide discovery and demonstrate the potential of integrating multiple protein modalities for more effective therapeutic peptide discovery.

Zengyi Wo, Minglai Shao 0001, Shiyu Zhang 0001, Ruijie Wang 0004

This paper presents the Local Homophily Graph Neural Network (LH-GNN), a novel framework for Graph Anomaly Detection (GAD). Anomalous activities in graphs often exhibit a complex interplay of homophily and heterophily, with our analysis revealing that anomalous nodes typically display a higher degree of heterophily compared to normal nodes. Existing GNN-based methods start to incorporate heterophily modeling but fail to address two critical challenges: (1) the efficiency challenge, as traditional spectral decomposition based methods are computationally expensive, and (2) the local homophily estimation challenge, where prior knowledge of node-wise homophily ratios is often unavailable. To address these challenges, LH-GNN introduces a lightweight polynomial graph filter that dynamically adjusts to node-specific homophily ratios, enabling efficient representation learning for both normal and anomalous nodes through adaptable heterophilic and homophilic bases. This design achieves linear time complexity, significantly improving computational efficiency. Additionally, we propose an iterative prototype learning strategy to estimate local homophily values without requiring additional labels. This strategy leverages class prototypes and uncertainty measures to assign reliable pseudo-labels, effectively capturing node-wise homophily. Together, these innovations enable LH-GNN to overcome the limitations of existing methods. Extensive experiments on four benchmark datasets demonstrate that LH-GNN outperforms state-of-the-art methods in both effectiveness and efficiency, achieving 4.4% improvements in detection accuracy and 11× computational speedup.

Zixuan Weng, Jindong Han, Wenzhao Jiang, Hao Liu 0026

In recent years, advancements in deep learning have spurred the development of numerous models for Long-term Time Series Forecasting (LTSF). However, most existing approaches struggle to fully capture the complex and structured dependencies inherent in time series data. In this work, we identify and formally define three critical dependencies that are fundamental to forecasting accuracy: order dependency and semantic dependency along the temporal dimension, as well as cross-variate dependency across the feature dimension. These dependencies are often treated in isolation, and improper handling can introduce noise and degrade forecasting performance. To bridge this gap, we investigate the potential of State Space Models (SSMs) for LTSF and emphasize their inherent advantages in capturing these essential dependencies. Additionally, we empirically observe that excessive nonlinearity in conventional SSMs introduce redundancy when applied to semantically sparse time series data. Motivated by this insight, we propose SDE (Simplified and Disentangled Dependency Encoding), a novel framework designed to enhance the capability of SSMs for LTSF. Specifically, we first eliminate unnecessary nonlinearities in vanilla SSMs, thereby improving the suitability for time series forecasting. Building on this foundation, we introduce a disentangled encoding strategy, which empowers SSMs to efficiently model cross-variate dependencies while mitigating interference between the temporal and feature dimensions. Furthermore, we provide rigorous theoretical justifications to substantiate our design choices. Extensive experiments on nine real-world benchmark datasets demonstrate that SDE-enhanced SSMs consistently outperform state-of-the-art time series forecasting models. Our code is available at https://github.com/YukinoAsuna/SAMBA.

Yi Wen 0001, Yue Liu 0008, Derong Xu, Huishi Luo, Pengyue Jia, Yiqing Wu, Siwei Wang 0001, Ke Liang 0006, Maolin Wang 0001, Yiqi Wang 0001 等

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.