论文分享 | 推荐系统相关研究进展

我们从2024-12-12到2024-12-30的39篇文章中精选出5篇优秀的工作分享给读者。

1. Whom do Explanations Serve? A Systematic Literature Survey of User Characteristics in Explainable Recommender Systems Evaluation
2. Muse: A Multimodal Conversational Recommendation Dataset with Scenario-Grounded User Profiles
3. HEC-GCN: Hypergraph Enhanced Cascading Graph Convolution Network for Multi-Behavior Recommendation
4. Molar: Multimodal LLMs with Collaborative Filtering Alignment for Enhanced Sequential Recommendation
5. Are Longer Prompts Always Better? Prompt Selection in Large Language Models for Recommendation Systems

1.Whom do Explanations Serve? A Systematic Literature Survey of User Characteristics in Explainable Recommender Systems Evaluation

Authors: Kathrin Wardatzky, Oana Inel, Luca Rossetto, Abraham Bernstein

https://arxiv.org/abs/2412.14193

论文摘要

Adding explanations to recommender systems is said to have multiple benefits, such as increasing user trust or system transparency. Previous work from other application areas suggests that specific user characteristics impact the users’ perception of the explanation. However, we rarely find this type of evaluation for recommender systems explanations. This paper addresses this gap by surveying 124 papers in which recommender systems explanations were evaluated in user studies. We analyzed their participant descriptions and study results where the impact of user characteristics on the explanation effects was measured. Our findings suggest that the results from the surveyed studies predominantly cover specific users who do not necessarily represent the users of recommender systems in the evaluation domain. This may seriously hamper the generalizability of any insights we may gain from current studies on explanations in recommender systems. We further find inconsistencies in the data reporting, which impacts the reproducibility of the reported results. Hence, we recommend actions to move toward a more inclusive and reproducible evaluation.

论文简评

该篇论文对解释推荐系统中用户特征评估的研究进行了综述，并提出了改进方法以提高研究的可靠性与普适性。首先，作者对解释推荐系统中用户特征的相关文献进行了全面而系统的调查，旨在填补现有研究中的空白。其次，通过分析这些研究，作者发现当前的研究存在一定的局限性，即缺乏对多样化用户群体的充分考虑以及数据报告的不一致性。为了解决这些问题，作者建议采取一系列措施来提升参与者的代表性、改善数据报告的标准和加强研究实践的可重复性。总的来说，这篇论文为解释推荐系统中如何更好地理解用户行为提供了宝贵的指导，有助于推动这一领域的发展。

2.Muse: A Multimodal Conversational Recommendation Dataset with Scenario-Grounded User Profiles

Authors: Zihan Wang, Xiaocui Yang, Yongkang Liu, Shi Feng, Daling Wang, Yifei Zhang

https://arxiv.org/abs/2412.18416

论文摘要

Current conversational recommendation systems focus predominantly on text. However, real-world recommendation settings are generally multimodal, causing a significant gap between existing research and practical applications.  To address this issue, we propose Muse, the first multimodal conversational recommendation dataset. Muse comprises 83,148 utterances from 7,000 conversations centered around the clothing domain.  Each conversation contains comprehensive multimodal interactions, rich elements, and natural dialogues.  Data in Muse are automatically synthesized by a multi-agent framework powered by multimodal large language models (MLLMs).  It innovatively derives user profiles from real-world scenarios rather than depending on manual design and historical data for better scalability, and then it fulfills conversation simulation and optimization.  Both human and LLM evaluations demonstrate the high quality of conversations in \textsc{Muse}. Additionally, fine-tuning experiments on three MLLMs demonstrate Muse's learnable patterns for recommendations and responses, confirming its value for multimodal conversational recommendation.  Our dataset and codes are available at https://anonymous.4open.science/r/Muse-0086.

论文简评

综上所述，本文《MUSE：一个聚焦于多模态交互的对话推荐数据集》提出了一个全新的研究课题——“多模态对话推荐系统”，旨在解决现有文本推荐系统的局限性，并通过跨领域合作，融合了语音、图像等多种信息源的数据，构建了一个具有丰富场景和真实用户行为特征的对话推荐模型。该研究不仅解决了现有研究中的技术瓶颈，还通过实证验证了其有效性，在提高机器学习模型的推荐能力方面发挥了重要作用。此外，文中对MUSE数据集进行了详细的介绍和分析，展示了它如何满足当前对话推荐领域的实际需求，为后续的研究提供了宝贵的经验和技术支持。总的来说，这篇文章是一个非常有潜力的研究成果，值得进一步深入探讨和应用。

3.HEC-GCN: Hypergraph Enhanced Cascading Graph Convolution Network for Multi-Behavior Recommendation

Authors: Yabo Yin, Xiaofei Zhu, Wenshan Wang, Yihao Zhang, Pengfei Wang, Yixing Fan, Jiafeng Guo

https://arxiv.org/abs/2412.14476

论文摘要

Multi-behavior recommendation (MBR) has garnered growing attention recently due to its ability to mitigate the sparsity issue by inferring user preferences from various auxiliary behaviors to improve predictions for the target behavior. Although existing research on MBR has yielded impressive results, they still face two major limitations. First, previous methods mainly focus on modeling f ine-grained interaction information between users and items under each behavior, which may suffer from sparsity issue. Second, existing models usually concentrate on exploiting dependencies between two consecutive behaviors, leaving intra- and inter-behavior consistency largely unexplored. To the end, we propose a novel approach named Hypergraph Enhanced Cascading Graph Convolution Network for multi-behavior recommendation (HEC-GCN). To be specific, we first explore both f ine- and coarse-grained correlations among users or items of each behavior by simultaneously modeling the behavior-specific interaction graph and its corresponding hypergraph in a cascaded manner. Then, we propose a behavior consistency-guided alignment strategy that ensures consistent representations between the interaction graph and its associated hypergraph for each behavior, while also maintaining representation consistency across different behaviors. Extensive experiments and analyses on three public benchmark datasets demonstrate that our proposed approach is consistently superior to previous state-of-the-art methods due to its capability to effectively attenuate the sparsity issue as well as preserve both intra- and inter-behavior consistencies. The code is available at https://github.com/marqu22/HEC-GCN.git.

论文简评

《HEC-GCN：一种超图增强级联图卷积神经网络》这篇论文通过提出HEC-GCN模型，旨在解决用户交互数据中的稀疏性问题，并提升多行为推荐系统的能力。论文引入了超图的概念来捕获用户-物品交互之间的相关性，并提出了一个基于一致性引导的对齐策略。实验结果表明，HEC-GCN在基准数据集上的表现优于当前最先进的方法，提供了强有力的实证支持。

该论文的关键在于其能够有效利用超图捕捉用户-物品交互中的细粒度和粗粒度信息，以应对数据稀疏性的挑战。此外，它还引入了一种新的维度——超图——来处理用户-物品交互，这为模型的有效性和鲁棒性带来了显著改进。实验结果显示，HEC-GCN不仅在性能上超过了现有的竞争方法，而且在评估指标上取得了更好的结果。这些成果证明了HEC-GCN模型具有很高的实用性，并且对于改善多行为推荐系统的性能具有重要意义。

4.Molar: Multimodal LLMs with Collaborative Filtering Alignment for Enhanced Sequential Recommendation

Authors: Yucong Luo, Qitao Qin, Hao Zhang, Mingyue Cheng, Ruiran Yan, Kefan Wang, Jie Ouyang

https://arxiv.org/abs/2412.18176

论文摘要

Sequential recommendation (SR) systems have evolved significantly over the past decade, tran sitioning from traditional collaborative filtering to deep learning approaches and, more recently, to large language models (LLMs). While the adoption of LLMs has driven substantial ad vancements, these models inherently lack col laborative filtering information, relying primar The User has watched North by Northwest Back to the Future Toy Story predict the next moviethe user will watch. North by Northwest Back to the Future Toy Story LLM Word Embedding Piror LLM-based Recommendation Item Text embedding Pre-Alignment LLM-based Recommender (a) Multimodal  Metadata Ours (b) North by  Northwest Back to  the Future Multimodal LLM Toy Story User Content embedding Post-Alignment Recommendation Layer  Text Metadata Traditional Rec Model Item ID embedding  Traditional Rec Model User ID embedding ID Number ID Number ily on textual content data neglecting other modalities and thus failing to achieve optimal recommendation performance. To address this limitation, we propose Molar, a Multimodal large language sequential recommendation framework that integrates multiple content modalities with ID information to capture col laborative signals effectively. Molar employs an MLLMtogenerate unified item representa tions from both textual and non-textual data, facilitating comprehensive multimodal mod eling and enriching item embeddings. Addi tionally, it incorporates collaborative filtering signals through a post-alignment mechanism, which aligns user representations from content based and ID-based models, ensuring precise personalization and robust performance. By seamlessly combining multimodal content with collaborative filtering insights, Molar captures both user interests and contextual semantics, leading to superior recommendation accuracy. Extensive experiments validate that Molar sig nificantly outperforms traditional and LLM based baselines, highlighting its strength in uti lizing multimodal data and collaborative sig nals for sequential recommendation tasks. The source code is available.

论文简评

本文研究了利用大规模语言模型对序列推荐系统中多模态内容数据进行集成的方法。提出了一个多模态物品表示模型（MIRM）和动态用户嵌入生成器（DUEG），以增强推荐精度，并通过大量实验验证证明其性能优于现有方法。该架构设计清晰，明确区分了MIRM和DUEG的角色。综上所述，该文提出了一种新颖而有效的解决方案，为解决当前推荐系统中的多模态问题提供了新的思路和方法。

5.Are Longer Prompts Always Better? Prompt Selection in Large Language Models for Recommendation Systems

Authors: Genki Kusano, Kosuke Akimoto, Kunihiro Takeoka

https://arxiv.org/abs/2412.14454

论文摘要

In large language models (LLM)-based recommendation systems (LLM-RSs), accurately predicting user preferences by leveraging the general knowledge of LLMs is possible without requiring extensive training data. By converting recommendation tasks into natural language inputs called prompts, LLM-RSs can efficiently address issues that have been difficult to resolve due to data scarcity but are crucial in applications such as cold-start and cross-domain problems. However, when applying this in practice, selecting the prompt that best matches the tasks and data is essential. Although numerous prompts have been proposed in LLM-RSs and representing the target user in prompts significantly impacts recommendation accuracy, there are still no clear guidelines for selecting specific prompts.

In this paper, we categorize and analyze prompts from previous research to establish practical prompt selection guidelines. Through 450 experiments with 90 prompts and five real-world datasets, we examined the relationship between prompts and dataset characteristics regarding recommendation accuracy. We found that no single prompt consistently outperforms others; thus, selecting prompts based on dataset characteristics is crucial. Here, we propose a prompt selection method that achieves higher accuracy with minimal validation data. Because increasing the number of prompts to explore raises costs, we also introduce a cost-efficient strategy using high-performance and cost-effective LLMs, significantly reducing exploration costs while maintaining high prediction accuracy. Our work offers valuable insights into prompt selection, advancing accurate and efficient LLM-RSs.

论文简评

这篇论文对大型语言模型驱动的推荐系统中的提示选择进行了深入研究。它根据之前的研究分类提示，并在五个数据集上进行了广泛的实验，以分析提示与数据特征之间的关系。研究表明，提示的选择对于最佳性能至关重要，因此提出了一种高效的提示选择方法，该方法平衡了准确性与成本。通过这个研究，论文解决了实际应用中的一个重要问题，并提供了丰富的实证数据来支持其结论。

我们欢迎您在评论区中留下宝贵的建议！包括但不限于：

• 可以提出推文中论文简评的不足！
• 可以分享最近更值得推荐的论文并给出理由！

推荐阅读