Question recommendation is a task that sequentially recommends questions for students to enhance their learning efficiency. That is, given the learning history and learning target of a student, a question recommender is supposed to select the question that will bring the most improvement for students. Previous methods typically model the question recommendation as a sequential decision-making problem, estimating students' learning state with the learning history, and feeding the learning state with the learning target to a neural network to select the recommended question from a question set. However, previous methods are faced with two challenges: (1) learning history is unavailable in the cold start scenario, which makes the recommender generate inappropriate recommendations; (2) the size of the question set is much large, which makes it difficult for the recommender to select the best question precisely. To address the challenges, we propose a method called hierarchical large language model for question recommendation (HierLLM), which is a LLM-based hierarchical structure. The LLM-based structure enables HierLLM to tackle the cold start issue with the strong reasoning abilities of LLM. The hierarchical structure takes advantage of the fact that the number of concepts is significantly smaller than the number of questions, narrowing the range of selectable questions by first identifying the relevant concept for the to-recommend question, and then selecting the recommended question based on that concept. This hierarchical structure reduces the difficulty of the recommendation.To investigate the performance of HierLLM, we conduct extensive experiments, and the results demonstrate the outstanding performance of HierLLM.
{"title":"HierLLM: Hierarchical Large Language Model for Question Recommendation","authors":"Yuxuan Liu, Haipeng Liu, Ting Long","doi":"arxiv-2409.06177","DOIUrl":"https://doi.org/arxiv-2409.06177","url":null,"abstract":"Question recommendation is a task that sequentially recommends questions for\u0000students to enhance their learning efficiency. That is, given the learning\u0000history and learning target of a student, a question recommender is supposed to\u0000select the question that will bring the most improvement for students. Previous\u0000methods typically model the question recommendation as a sequential\u0000decision-making problem, estimating students' learning state with the learning\u0000history, and feeding the learning state with the learning target to a neural\u0000network to select the recommended question from a question set. However,\u0000previous methods are faced with two challenges: (1) learning history is\u0000unavailable in the cold start scenario, which makes the recommender generate\u0000inappropriate recommendations; (2) the size of the question set is much large,\u0000which makes it difficult for the recommender to select the best question\u0000precisely. To address the challenges, we propose a method called hierarchical\u0000large language model for question recommendation (HierLLM), which is a\u0000LLM-based hierarchical structure. The LLM-based structure enables HierLLM to\u0000tackle the cold start issue with the strong reasoning abilities of LLM. The\u0000hierarchical structure takes advantage of the fact that the number of concepts\u0000is significantly smaller than the number of questions, narrowing the range of\u0000selectable questions by first identifying the relevant concept for the\u0000to-recommend question, and then selecting the recommended question based on\u0000that concept. This hierarchical structure reduces the difficulty of the\u0000recommendation.To investigate the performance of HierLLM, we conduct extensive\u0000experiments, and the results demonstrate the outstanding performance of\u0000HierLLM.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"7 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205327","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A good medical ontology is expected to cover its domain completely and correctly. On the other hand, large ontologies are hard to build, hard to understand, and hard to maintain. Thus, adding new concepts (often multi-word concepts) to an existing ontology must be done judiciously. Only "good" concepts should be added; however, it is difficult to define what makes a concept good. In this research, we propose a metric to measure the goodness of a concept. We identified factors that appear to influence goodness judgments of medical experts and combined them into a single metric. These factors include concept name length (in words), concept occurrence frequency in the medical literature, and syntactic categories of component words. As an added factor, we used the simplicity of a term after mapping it into a specific foreign language. We performed Bayesian optimization of factor weights to achieve maximum agreement between the metric and three medical experts. The results showed that our metric had a 50.67% overall agreement with the experts, as measured by Krippendorff's alpha.
{"title":"What makes a good concept anyway ?","authors":"Naren Khatwani, James Geller","doi":"arxiv-2409.06150","DOIUrl":"https://doi.org/arxiv-2409.06150","url":null,"abstract":"A good medical ontology is expected to cover its domain completely and\u0000correctly. On the other hand, large ontologies are hard to build, hard to\u0000understand, and hard to maintain. Thus, adding new concepts (often multi-word\u0000concepts) to an existing ontology must be done judiciously. Only \"good\"\u0000concepts should be added; however, it is difficult to define what makes a\u0000concept good. In this research, we propose a metric to measure the goodness of\u0000a concept. We identified factors that appear to influence goodness judgments of\u0000medical experts and combined them into a single metric. These factors include\u0000concept name length (in words), concept occurrence frequency in the medical\u0000literature, and syntactic categories of component words. As an added factor, we\u0000used the simplicity of a term after mapping it into a specific foreign\u0000language. We performed Bayesian optimization of factor weights to achieve\u0000maximum agreement between the metric and three medical experts. The results\u0000showed that our metric had a 50.67% overall agreement with the experts, as\u0000measured by Krippendorff's alpha.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"6 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205328","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Practitioners working on dense retrieval today face a bewildering number of choices. Beyond selecting the embedding model, another consequential choice is the actual implementation of nearest-neighbor vector search. While best practices recommend HNSW indexes, flat vector indexes with brute-force search represent another viable option, particularly for smaller corpora and for rapid prototyping. In this paper, we provide experimental results on the BEIR dataset using the open-source Lucene search library that explicate the tradeoffs between HNSW and flat indexes (including quantized variants) from the perspectives of indexing time, query evaluation performance, and retrieval quality. With additional comparisons between dense and sparse retrievers, our results provide guidance for today's search practitioner in understanding the design space of dense and sparse retrievers. To our knowledge, we are the first to provide operational advice supported by empirical experiments in this regard.
{"title":"Operational Advice for Dense and Sparse Retrievers: HNSW, Flat, or Inverted Indexes?","authors":"Jimmy Lin","doi":"arxiv-2409.06464","DOIUrl":"https://doi.org/arxiv-2409.06464","url":null,"abstract":"Practitioners working on dense retrieval today face a bewildering number of\u0000choices. Beyond selecting the embedding model, another consequential choice is\u0000the actual implementation of nearest-neighbor vector search. While best\u0000practices recommend HNSW indexes, flat vector indexes with brute-force search\u0000represent another viable option, particularly for smaller corpora and for rapid\u0000prototyping. In this paper, we provide experimental results on the BEIR dataset\u0000using the open-source Lucene search library that explicate the tradeoffs\u0000between HNSW and flat indexes (including quantized variants) from the\u0000perspectives of indexing time, query evaluation performance, and retrieval\u0000quality. With additional comparisons between dense and sparse retrievers, our\u0000results provide guidance for today's search practitioner in understanding the\u0000design space of dense and sparse retrievers. To our knowledge, we are the first\u0000to provide operational advice supported by empirical experiments in this\u0000regard.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"14 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yongsu Ahn, Quinn K Wolter, Jonilyn Dick, Janet Dick, Yu-Ru Lin
Recommender systems have become integral to digital experiences, shaping user interactions and preferences across various platforms. Despite their widespread use, these systems often suffer from algorithmic biases that can lead to unfair and unsatisfactory user experiences. This study introduces an interactive tool designed to help users comprehend and explore the impacts of algorithmic harms in recommender systems. By leveraging visualizations, counterfactual explanations, and interactive modules, the tool allows users to investigate how biases such as miscalibration, stereotypes, and filter bubbles affect their recommendations. Informed by in-depth user interviews, this tool benefits both general users and researchers by increasing transparency and offering personalized impact assessments, ultimately fostering a better understanding of algorithmic biases and contributing to more equitable recommendation outcomes. This work provides valuable insights for future research and practical applications in mitigating bias and enhancing fairness in machine learning algorithms.
{"title":"Interactive Counterfactual Exploration of Algorithmic Harms in Recommender Systems","authors":"Yongsu Ahn, Quinn K Wolter, Jonilyn Dick, Janet Dick, Yu-Ru Lin","doi":"arxiv-2409.06916","DOIUrl":"https://doi.org/arxiv-2409.06916","url":null,"abstract":"Recommender systems have become integral to digital experiences, shaping user\u0000interactions and preferences across various platforms. Despite their widespread\u0000use, these systems often suffer from algorithmic biases that can lead to unfair\u0000and unsatisfactory user experiences. This study introduces an interactive tool\u0000designed to help users comprehend and explore the impacts of algorithmic harms\u0000in recommender systems. By leveraging visualizations, counterfactual\u0000explanations, and interactive modules, the tool allows users to investigate how\u0000biases such as miscalibration, stereotypes, and filter bubbles affect their\u0000recommendations. Informed by in-depth user interviews, this tool benefits both\u0000general users and researchers by increasing transparency and offering\u0000personalized impact assessments, ultimately fostering a better understanding of\u0000algorithmic biases and contributing to more equitable recommendation outcomes.\u0000This work provides valuable insights for future research and practical\u0000applications in mitigating bias and enhancing fairness in machine learning\u0000algorithms.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"2 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205324","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Weicong Qin, Yi Xu, Weijie Yu, Chenglei Shen, Xiao Zhang, Ming He, Jianping Fan, Jun Xu
Sequence recommendation (SeqRec) aims to predict the next item a user will interact with by understanding user intentions and leveraging collaborative filtering information. Large language models (LLMs) have shown great promise in recommendation tasks through prompt-based, fixed reflection libraries, and fine-tuning techniques. However, these methods face challenges, including lack of supervision, inability to optimize reflection sources, inflexibility to diverse user needs, and high computational costs. Despite promising results, current studies primarily focus on reflections of users' explicit preferences (e.g., item titles) while neglecting implicit preferences (e.g., brands) and collaborative filtering information. This oversight hinders the capture of preference shifts and dynamic user behaviors. Additionally, existing approaches lack mechanisms for reflection evaluation and iteration, often leading to suboptimal recommendations. To address these issues, we propose the Mixture of REflectors (MoRE) framework, designed to model and learn dynamic user preferences in SeqRec. Specifically, MoRE introduces three reflectors for generating LLM-based reflections on explicit preferences, implicit preferences, and collaborative signals. Each reflector incorporates a self-improving strategy, termed refining-and-iteration, to evaluate and iteratively update reflections. Furthermore, a meta-reflector employs a contextual bandit algorithm to select the most suitable expert and corresponding reflections for each user's recommendation, effectively capturing dynamic preferences. Extensive experiments on three real-world datasets demonstrate that MoRE consistently outperforms state-of-the-art methods, requiring less training time and GPU memory compared to other LLM-based approaches in SeqRec.
{"title":"Enhancing Sequential Recommendations through Multi-Perspective Reflections and Iteration","authors":"Weicong Qin, Yi Xu, Weijie Yu, Chenglei Shen, Xiao Zhang, Ming He, Jianping Fan, Jun Xu","doi":"arxiv-2409.06377","DOIUrl":"https://doi.org/arxiv-2409.06377","url":null,"abstract":"Sequence recommendation (SeqRec) aims to predict the next item a user will\u0000interact with by understanding user intentions and leveraging collaborative\u0000filtering information. Large language models (LLMs) have shown great promise in\u0000recommendation tasks through prompt-based, fixed reflection libraries, and\u0000fine-tuning techniques. However, these methods face challenges, including lack\u0000of supervision, inability to optimize reflection sources, inflexibility to\u0000diverse user needs, and high computational costs. Despite promising results,\u0000current studies primarily focus on reflections of users' explicit preferences\u0000(e.g., item titles) while neglecting implicit preferences (e.g., brands) and\u0000collaborative filtering information. This oversight hinders the capture of\u0000preference shifts and dynamic user behaviors. Additionally, existing approaches\u0000lack mechanisms for reflection evaluation and iteration, often leading to\u0000suboptimal recommendations. To address these issues, we propose the Mixture of\u0000REflectors (MoRE) framework, designed to model and learn dynamic user\u0000preferences in SeqRec. Specifically, MoRE introduces three reflectors for\u0000generating LLM-based reflections on explicit preferences, implicit preferences,\u0000and collaborative signals. Each reflector incorporates a self-improving\u0000strategy, termed refining-and-iteration, to evaluate and iteratively update\u0000reflections. Furthermore, a meta-reflector employs a contextual bandit\u0000algorithm to select the most suitable expert and corresponding reflections for\u0000each user's recommendation, effectively capturing dynamic preferences.\u0000Extensive experiments on three real-world datasets demonstrate that MoRE\u0000consistently outperforms state-of-the-art methods, requiring less training time\u0000and GPU memory compared to other LLM-based approaches in SeqRec.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"117 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205326","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Qitao Qin, Yucong Luo, Mingyue Cheng, Qingyang Mao, Chenyi Lei
Federated recommendation (FedRec) preserves user privacy by enabling decentralized training of personalized models, but this architecture is inherently vulnerable to adversarial attacks. Significant research has been conducted on targeted attacks in FedRec systems, motivated by commercial and social influence considerations. However, much of this work has largely overlooked the differential robustness of recommendation models. Moreover, our empirical findings indicate that existing targeted attack methods achieve only limited effectiveness in Federated Sequential Recommendation (FSR) tasks. Driven by these observations, we focus on investigating targeted attacks in FSR and propose a novel dualview attack framework, named DV-FSR. This attack method uniquely combines a sampling-based explicit strategy with a contrastive learning-based implicit gradient strategy to orchestrate a coordinated attack. Additionally, we introduce a specific defense mechanism tailored for targeted attacks in FSR, aiming to evaluate the mitigation effects of the attack method we proposed. Extensive experiments validate the effectiveness of our proposed approach on representative sequential models.
{"title":"DV-FSR: A Dual-View Target Attack Framework for Federated Sequential Recommendation","authors":"Qitao Qin, Yucong Luo, Mingyue Cheng, Qingyang Mao, Chenyi Lei","doi":"arxiv-2409.07500","DOIUrl":"https://doi.org/arxiv-2409.07500","url":null,"abstract":"Federated recommendation (FedRec) preserves user privacy by enabling\u0000decentralized training of personalized models, but this architecture is\u0000inherently vulnerable to adversarial attacks. Significant research has been\u0000conducted on targeted attacks in FedRec systems, motivated by commercial and\u0000social influence considerations. However, much of this work has largely\u0000overlooked the differential robustness of recommendation models. Moreover, our\u0000empirical findings indicate that existing targeted attack methods achieve only\u0000limited effectiveness in Federated Sequential Recommendation (FSR) tasks.\u0000Driven by these observations, we focus on investigating targeted attacks in FSR\u0000and propose a novel dualview attack framework, named DV-FSR. This attack method\u0000uniquely combines a sampling-based explicit strategy with a contrastive\u0000learning-based implicit gradient strategy to orchestrate a coordinated attack.\u0000Additionally, we introduce a specific defense mechanism tailored for targeted\u0000attacks in FSR, aiming to evaluate the mitigation effects of the attack method\u0000we proposed. Extensive experiments validate the effectiveness of our proposed\u0000approach on representative sequential models.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205314","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The scale-space method is a well-established framework that constructs a hierarchical representation of an input signal and facilitates coarse-to-fine visual reasoning. Considering the terrain elevation function as the input signal, the scale-space method can identify and track significant topographic features across different scales. The number of scales a feature persists, called its life span, indicates the importance of that feature. In this way, important topographic features of a landscape can be selected, which are useful for many applications, including cartography, nautical charting, and land-use planning. The scale-space methods developed for terrain data use gridded Digital Elevation Models (DEMs) to represent the terrain. However, gridded DEMs lack the flexibility to adapt to the irregular distribution of input data and the varied topological complexity of different regions. Instead, Triangulated Irregular Networks (TINs) can be directly generated from irregularly distributed point clouds and accurately preserve important features. In this work, we introduce a novel scale-space analysis pipeline for TINs, addressing the multiple challenges in extending grid-based scale-space methods to TINs. Our pipeline can efficiently identify and track topologically important features on TINs. Moreover, it is capable of analyzing terrains with irregular boundaries, which poses challenges for grid-based methods. Comprehensive experiments show that, compared to grid-based methods, our TIN-based pipeline is more efficient, accurate, and has better resolution robustness.
尺度空间法是一种成熟的框架，它能构建输入信号的层次表示法，便于进行从粗到细的视觉推理。将地形高程函数视为输入信号，尺度空间法可以识别和跟踪不同尺度上的重要地形特征。地貌特征持续存在的尺度数（称为其寿命）表明了该特征的重要性。通过这种方法，可以筛选出景观中重要的地形特征，这在制图、海图绘制和土地利用规划等许多应用中都非常有用。为地形数据开发的比例空间方法使用网格数字高程模型（DEM）来表示地形。然而，网格数字高程模型缺乏灵活性，无法适应输入数据的不规则分布和不同地区的不同地形复杂性。相反，三角不规则网络（TIN）可以直接从不规则分布的点云生成，并准确地保留重要特征。在这项工作中，我们为 TINs 引入了一个新颖的尺度空间分析管道，解决了将基于网格的尺度空间方法扩展到 TINs 时所面临的多重挑战。此外，它还能分析具有不规则边界的地形，这对基于网格的方法构成了挑战。综合实验表明，与基于网格的方法相比，我们基于 TIN 的管道更高效、更准确，并且具有更好的分辨率鲁棒性。
{"title":"Critical Features Tracking on Triangulated Irregular Networks by a Scale-Space Method","authors":"Haoan Feng, Yunting Song, Leila De Floriani","doi":"arxiv-2409.06638","DOIUrl":"https://doi.org/arxiv-2409.06638","url":null,"abstract":"The scale-space method is a well-established framework that constructs a\u0000hierarchical representation of an input signal and facilitates coarse-to-fine\u0000visual reasoning. Considering the terrain elevation function as the input\u0000signal, the scale-space method can identify and track significant topographic\u0000features across different scales. The number of scales a feature persists,\u0000called its life span, indicates the importance of that feature. In this way,\u0000important topographic features of a landscape can be selected, which are useful\u0000for many applications, including cartography, nautical charting, and land-use\u0000planning. The scale-space methods developed for terrain data use gridded\u0000Digital Elevation Models (DEMs) to represent the terrain. However, gridded DEMs\u0000lack the flexibility to adapt to the irregular distribution of input data and\u0000the varied topological complexity of different regions. Instead, Triangulated\u0000Irregular Networks (TINs) can be directly generated from irregularly\u0000distributed point clouds and accurately preserve important features. In this\u0000work, we introduce a novel scale-space analysis pipeline for TINs, addressing\u0000the multiple challenges in extending grid-based scale-space methods to TINs.\u0000Our pipeline can efficiently identify and track topologically important\u0000features on TINs. Moreover, it is capable of analyzing terrains with irregular\u0000boundaries, which poses challenges for grid-based methods. Comprehensive\u0000experiments show that, compared to grid-based methods, our TIN-based pipeline\u0000is more efficient, accurate, and has better resolution robustness.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"16 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Tri Kurniawan Wijaya, Edoardo D'Amico, Gabor Fodor, Manuel V. Loureiro
Rs4rs is a web application designed to perform semantic search on recent papers from top conferences and journals related to Recommender Systems. Current scholarly search engine tools like Google Scholar, Semantic Scholar, and ResearchGate often yield broad results that fail to target the most relevant high-quality publications. Moreover, manually visiting individual conference and journal websites is a time-consuming process that primarily supports only syntactic searches. Rs4rs addresses these issues by providing a user-friendly platform where researchers can input their topic of interest and receive a list of recent, relevant papers from top Recommender Systems venues. Utilizing semantic search techniques, Rs4rs ensures that the search results are not only precise and relevant but also comprehensive, capturing papers regardless of variations in wording. This tool significantly enhances research efficiency and accuracy, thereby benefitting the research community and public by facilitating access to high-quality, pertinent academic resources in the field of Recommender Systems. Rs4rs is available at https://rs4rs.com.
Rs4rs 是一款网络应用程序，旨在对与推荐系统相关的顶级会议和期刊的最新论文进行语义搜索。目前的学术搜索引擎工具（如 Google Scholar、Semantic Scholar 和 ResearchGate）通常会产生广泛的搜索结果，无法锁定最相关的高质量出版物。此外，手动访问各个会议和期刊网站也是一个耗时的过程，而且主要只支持句法搜索。Rs4rs 利用语义搜索技术，确保搜索结果不仅精确、相关，而且全面，无论措辞如何变化，都能捕捉到论文。该工具大大提高了研究效率和准确性，从而为研究界和公众获取推荐系统领域高质量的相关学术资源提供了便利。Rs4rs可在https://rs4rs.com。
{"title":"Rs4rs: Semantically Find Recent Publications from Top Recommendation System-Related Venues","authors":"Tri Kurniawan Wijaya, Edoardo D'Amico, Gabor Fodor, Manuel V. Loureiro","doi":"arxiv-2409.05570","DOIUrl":"https://doi.org/arxiv-2409.05570","url":null,"abstract":"Rs4rs is a web application designed to perform semantic search on recent\u0000papers from top conferences and journals related to Recommender Systems.\u0000Current scholarly search engine tools like Google Scholar, Semantic Scholar,\u0000and ResearchGate often yield broad results that fail to target the most\u0000relevant high-quality publications. Moreover, manually visiting individual\u0000conference and journal websites is a time-consuming process that primarily\u0000supports only syntactic searches. Rs4rs addresses these issues by providing a\u0000user-friendly platform where researchers can input their topic of interest and\u0000receive a list of recent, relevant papers from top Recommender Systems venues.\u0000Utilizing semantic search techniques, Rs4rs ensures that the search results are\u0000not only precise and relevant but also comprehensive, capturing papers\u0000regardless of variations in wording. This tool significantly enhances research\u0000efficiency and accuracy, thereby benefitting the research community and public\u0000by facilitating access to high-quality, pertinent academic resources in the\u0000field of Recommender Systems. Rs4rs is available at https://rs4rs.com.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"17 8 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Bowen Zheng, Junjie Zhang, Hongyu Lu, Yu Chen, Ming Chen, Wayne Xin Zhao, Ji-Rong Wen
Graph neural network (GNN) has been a powerful approach in collaborative filtering (CF) due to its ability to model high-order user-item relationships. Recently, to alleviate the data sparsity and enhance representation learning, many efforts have been conducted to integrate contrastive learning (CL) with GNNs. Despite the promising improvements, the contrastive view generation based on structure and representation perturbations in existing methods potentially disrupts the collaborative information in contrastive views, resulting in limited effectiveness of positive alignment. To overcome this issue, we propose CoGCL, a novel framework that aims to enhance graph contrastive learning by constructing contrastive views with stronger collaborative information via discrete codes. The core idea is to map users and items into discrete codes rich in collaborative information for reliable and informative contrastive view generation. To this end, we initially introduce a multi-level vector quantizer in an end-to-end manner to quantize user and item representations into discrete codes. Based on these discrete codes, we enhance the collaborative information of contrastive views by considering neighborhood structure and semantic relevance respectively. For neighborhood structure, we propose virtual neighbor augmentation by treating discrete codes as virtual neighbors, which expands an observed user-item interaction into multiple edges involving discrete codes. Regarding semantic relevance, we identify similar users/items based on shared discrete codes and interaction targets to generate the semantically relevant view. Through these strategies, we construct contrastive views with stronger collaborative information and develop a triple-view graph contrastive learning approach. Extensive experiments on four public datasets demonstrate the effectiveness of our proposed approach.
{"title":"Enhancing Graph Contrastive Learning with Reliable and Informative Augmentation for Recommendation","authors":"Bowen Zheng, Junjie Zhang, Hongyu Lu, Yu Chen, Ming Chen, Wayne Xin Zhao, Ji-Rong Wen","doi":"arxiv-2409.05633","DOIUrl":"https://doi.org/arxiv-2409.05633","url":null,"abstract":"Graph neural network (GNN) has been a powerful approach in collaborative\u0000filtering (CF) due to its ability to model high-order user-item relationships.\u0000Recently, to alleviate the data sparsity and enhance representation learning,\u0000many efforts have been conducted to integrate contrastive learning (CL) with\u0000GNNs. Despite the promising improvements, the contrastive view generation based\u0000on structure and representation perturbations in existing methods potentially\u0000disrupts the collaborative information in contrastive views, resulting in\u0000limited effectiveness of positive alignment. To overcome this issue, we propose\u0000CoGCL, a novel framework that aims to enhance graph contrastive learning by\u0000constructing contrastive views with stronger collaborative information via\u0000discrete codes. The core idea is to map users and items into discrete codes\u0000rich in collaborative information for reliable and informative contrastive view\u0000generation. To this end, we initially introduce a multi-level vector quantizer\u0000in an end-to-end manner to quantize user and item representations into discrete\u0000codes. Based on these discrete codes, we enhance the collaborative information\u0000of contrastive views by considering neighborhood structure and semantic\u0000relevance respectively. For neighborhood structure, we propose virtual neighbor\u0000augmentation by treating discrete codes as virtual neighbors, which expands an\u0000observed user-item interaction into multiple edges involving discrete codes.\u0000Regarding semantic relevance, we identify similar users/items based on shared\u0000discrete codes and interaction targets to generate the semantically relevant\u0000view. Through these strategies, we construct contrastive views with stronger\u0000collaborative information and develop a triple-view graph contrastive learning\u0000approach. Extensive experiments on four public datasets demonstrate the\u0000effectiveness of our proposed approach.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"55 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205353","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Arkadeep Acharya, Rudra Murthy, Vishwajeet Kumar, Jaydeep Sen
Despite significant progress in multilingual information retrieval, the lack of models capable of effectively supporting multiple languages, particularly low-resource like Indic languages, remains a critical challenge. This paper presents NLLB-E5: A Scalable Multilingual Retrieval Model. NLLB-E5 leverages the in-built multilingual capabilities in the NLLB encoder for translation tasks. It proposes a distillation approach from multilingual retriever E5 to provide a zero-shot retrieval approach handling multiple languages, including all major Indic languages, without requiring multilingual training data. We evaluate the model on a comprehensive suite of existing benchmarks, including Hindi-BEIR, highlighting its robust performance across diverse languages and tasks. Our findings uncover task and domain-specific challenges, providing valuable insights into the retrieval performance, especially for low-resource languages. NLLB-E5 addresses the urgent need for an inclusive, scalable, and language-agnostic text retrieval model, advancing the field of multilingual information access and promoting digital inclusivity for millions of users globally.
{"title":"NLLB-E5: A Scalable Multilingual Retrieval Model","authors":"Arkadeep Acharya, Rudra Murthy, Vishwajeet Kumar, Jaydeep Sen","doi":"arxiv-2409.05401","DOIUrl":"https://doi.org/arxiv-2409.05401","url":null,"abstract":"Despite significant progress in multilingual information retrieval, the lack\u0000of models capable of effectively supporting multiple languages, particularly\u0000low-resource like Indic languages, remains a critical challenge. This paper\u0000presents NLLB-E5: A Scalable Multilingual Retrieval Model. NLLB-E5 leverages\u0000the in-built multilingual capabilities in the NLLB encoder for translation\u0000tasks. It proposes a distillation approach from multilingual retriever E5 to\u0000provide a zero-shot retrieval approach handling multiple languages, including\u0000all major Indic languages, without requiring multilingual training data. We\u0000evaluate the model on a comprehensive suite of existing benchmarks, including\u0000Hindi-BEIR, highlighting its robust performance across diverse languages and\u0000tasks. Our findings uncover task and domain-specific challenges, providing\u0000valuable insights into the retrieval performance, especially for low-resource\u0000languages. NLLB-E5 addresses the urgent need for an inclusive, scalable, and\u0000language-agnostic text retrieval model, advancing the field of multilingual\u0000information access and promoting digital inclusivity for millions of users\u0000globally.","PeriodicalId":501281,"journal":{"name":"arXiv - CS - Information Retrieval","volume":"63 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142205393","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：