recsys2023.md

RECSYS2023 Paper List

论文	作者	组织	摘要	翻译	代码	引用数
TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation	Keqin Bao, Jizhi Zhang, Yang Zhang, Wenjie Wang, Fuli Feng, Xiangnan He				code	3
Is ChatGPT Fair for Recommendation? Evaluating Fairness in Large Language Model Recommendation	Jizhi Zhang, Keqin Bao, Yang Zhang, Wenjie Wang, Fuli Feng, Xiangnan He				code	2
On the Consistency, Discriminative Power and Robustness of Sampled Metrics in Offline Top-N Recommender System Evaluation	Yang Liu, Alan Medlar, Dorota Glowacka				code	1
Large Language Model Augmented Narrative Driven Recommendations	Sheshera Mysore, Andrew McCallum, Hamed Zamani				code	1
AdaptEx: A Self-Service Contextual Bandit Platform	William Black, Ercument Ilhan, Andrea Marchini, Vilda Markeviciute				code	1
Learning from Negative User Feedback and Measuring Responsiveness for Sequential Recommenders	Yueqi Wang, Yoni Halpern, Shuo Chang, Jingchen Feng, Elaine Ya Le, Longfei Li, Xujian Liang, MinCheng Huang, Shane Li, Alex Beutel, Yaping Zhang, Shuchao Bi	Google, Mountain View, CA 94043 USA	Sequential recommenders have been widely used in industry due to their strength in modeling user preferences. While these models excel at learning a user's positive interests, less attention has been paid to learning from negative user feedback. Negative user feedback is an important lever of user control, and comes with an expectation that recommenders should respond quickly and reduce similar recommendations to the user. However, negative feedback signals are often ignored in the training objective of sequential retrieval models, which primarily aim at predicting positive user interactions. In this work, we incorporate explicit and implicit negative user feedback into the training objective of sequential recommenders in the retrieval stage using a "not-to-recommend" loss function that optimizes for the log-likelihood of not recommending items with negative feedback. We demonstrate the effectiveness of this approach using live experiments on a large-scale industrial recommender system. Furthermore, we address a challenge in measuring recommender responsiveness to negative feedback by developing a counterfactual simulation framework to compare recommender responses between different user actions, showing improved responsiveness from the modeling change.	由于序列推荐器在建模用户偏好方面的优势，它在工业中得到了广泛的应用。虽然这些模型擅长于学习用户的积极兴趣，但很少有人注意从消极的用户反馈中学习。负面的用户反馈是用户控制的一个重要杠杆，并伴随着一个期望，即推荐者应该快速响应并减少对用户的类似推荐。然而，在序贯检索模型的训练目标中，负反馈信号往往被忽略，而序贯检索模型的训练目标主要是预测正向用户交互。在这项工作中，我们将显性和隐性的负面用户反馈纳入顺序推荐者在检索阶段的训练目标中，使用一个“不推荐”的损失函数，该函数优化了不推荐负面反馈项目的对数可能性。我们通过在大规模工业推荐系统上的实验证明了这种方法的有效性。此外，我们通过开发一个反事实模拟框架来比较不同用户行为之间的推荐响应，从而解决了测量推荐响应负面反馈的挑战，显示了来自建模更改的更好响应。	code	0
gSASRec: Reducing Overconfidence in Sequential Recommendation Trained with Negative Sampling	Aleksandr Vladimirovich Petrov, Craig MacDonald	School of Computing Science, University of Glasgow, United Kingdom; University of Glasgow, United Kingdom	A large catalogue size is one of the central challenges in training recommendation models: a large number of items makes them memory and computationally inefficient to compute scores for all items during training, forcing these models to deploy negative sampling. However, negative sampling increases the proportion of positive interactions in the training data, and therefore models trained with negative sampling tend to overestimate the probabilities of positive interactions a phenomenon we call overconfidence. While the absolute values of the predicted scores or probabilities are not important for the ranking of retrieved recommendations, overconfident models may fail to estimate nuanced differences in the top-ranked items, resulting in degraded performance. In this paper, we show that overconfidence explains why the popular SASRec model underperforms when compared to BERT4Rec. This is contrary to the BERT4Rec authors explanation that the difference in performance is due to the bi-directional attention mechanism. To mitigate overconfidence, we propose a novel Generalised Binary Cross-Entropy Loss function (gBCE) and theoretically prove that it can mitigate overconfidence. We further propose the gSASRec model, an improvement over SASRec that deploys an increased number of negatives and the gBCE loss. We show through detailed experiments on three datasets that gSASRec does not exhibit the overconfidence problem. As a result, gSASRec can outperform BERT4Rec (e.g. +9.47% NDCG on the MovieLens-1M dataset), while requiring less training time (e.g. -73% training time on MovieLens-1M). Moreover, in contrast to BERT4Rec, gSASRec is suitable for large datasets that contain more than 1 million items.	大目录规模是培训推荐模型的核心挑战之一: 大量的项目使得它们在计算培训期间所有项目的分数时内存和计算效率低下，迫使这些模型部署负抽样。然而，负抽样增加了训练数据中正相互作用的比例，因此用负抽样训练的模型倾向于高估正相互作用的概率，我们称之为过度自信现象。虽然预测分数或概率的绝对值对检索推荐的排名并不重要，但过度自信的模型可能无法估计排名最高的项目的细微差异，导致性能下降。在本文中，我们表明，过度自信解释了为什么流行的 SASRec 模型表现不如 BERT4Rec。这与 BERT4Rec 的作者解释的性能差异是由于双向注意机制相反。为了减轻过度自信，我们提出了一种新的广义二元交叉熵损失函数(gBCE) ，并从理论上证明了它可以减轻过度自信。我们进一步提出了 gSASRec 模型，这是对 SASRec 模型的一个改进，它部署了更多的负片和 gBCE 损失。通过对三个数据集的详细实验，我们发现 gSASRec 不存在过度自信问题。因此，gSASRec 的性能优于 BERT4Rec (例如，在 MovieLens-1M 数据集上 + 9.47% NDCG) ，同时需要较少的训练时间(例如，在 MovieLens-1M 上 -73% 的训练时间)。此外，与 BERT4Rec 不同，gSASRec 适用于包含超过100万个项目的大型数据集。	code	0
Rethinking Multi-Interest Learning for Candidate Matching in Recommender Systems	Yueqi Xie, Jingqi Gao, Peilin Zhou, Qichen Ye, Yining Hua, Jae Boum Kim, Fangzhao Wu, Sunghun Kim	HKUST, Hong Kong, Peoples R China; MSRA, Beijing, Peoples R China; MIT, Cambridge, MA 02139 USA; Peking Univ, Beijing, Peoples R China; Upstage, Hong Kong, Peoples R China; HKUST gz, Hong Kong, Peoples R China	Existing research efforts for multi-interest candidate matching in recommender systems mainly focus on improving model architecture or incorporating additional information, neglecting the importance of training schemes. This work revisits the training framework and uncovers two major problems hindering the expressiveness of learned multi-interest representations. First, the current training objective (i.e., uniformly sampled softmax) fails to effectively train discriminative representations in a multi-interest learning scenario due to the severe increase in easy negative samples. Second, a routing collapse problem is observed where each learned interest may collapse to express information only from a single item, resulting in information loss. To address these issues, we propose the REMI framework, consisting of an Interest-aware Hard Negative mining strategy (IHN) and a Routing Regularization (RR) method. IHN emphasizes interest-aware hard negatives by proposing an ideal sampling distribution and developing a Monte-Carlo strategy for efficient approximation. RR prevents routing collapse by introducing a novel regularization term on the item-to-interest routing matrices. These two components enhance the learned multi-interest representations from both the optimization objective and the composition information. REMI is a general framework that can be readily applied to various existing multi-interest candidate matching methods. Experiments on three real-world datasets show our method can significantly improve state-of-the-art methods with easy implementation and negligible computational overhead. The source code is available at https://github.com/Tokkiu/REMI.	推荐系统中多兴趣候选人匹配的研究主要集中在改进模型结构或引入额外信息上，忽视了培训方案的重要性。这项工作重新审视了训练框架，发现了两个主要问题，阻碍了学习的多重兴趣表征的表达。首先，当前的训练目标(即均匀采样的软最大值)不能有效地训练多兴趣学习场景中的区分性表示，因为容易出现负样本的严重增加。其次，观察到一个路由折叠问题，其中每个学习兴趣可能会折叠成只表达单个项目的信息，从而导致信息损失。为了解决这些问题，我们提出了 REMI 框架，包括一个感兴趣的硬负面挖掘策略(IHN)和一个路由正则化(RR)方法。IHN 强调感兴趣的硬负面提出了一个理想的抽样分布和发展蒙特卡罗策略的有效逼近。RR 通过在项目感兴趣的路由矩阵上引入一个新的正则化项来防止路由崩溃。这两个部分从优化目标和组合信息两个方面增强了学习到的多兴趣表示。REMI 是一个通用框架，可以很容易地应用于各种现有的多兴趣候选匹配方法。在三个实际数据集上的实验表明，该方法可以显著改善最先进的方法，并且易于实现，计算开销可以忽略。源代码可在 https://github.com/tokkiu/remi 下载。	code	0
Understanding and Modeling Passive-Negative Feedback for Short-video Sequential Recommendation	Yunzhu Pan, Chen Gao, Jianxin Chang, Yanan Niu, Yang Song, Kun Gai, Depeng Jin, Yong Li	Beijing Kuaishou Technol Co Ltd, Beijing, Peoples R China; Unaffiliated, Beijing, Peoples R China; Univ Elect Sci & Technol China, Chengdu, Peoples R China; Tsinghua Univ, Beijing Natl Res Ctr Informat Sci & Technol, Dept Elect Engn, Beijing, Peoples R China	Sequential recommendation is one of the most important tasks in recommender systems, which aims to recommend the next interacted item with historical behaviors as input. Traditional sequential recommendation always mainly considers the collected positive feedback such as click, purchase, etc. However, in short-video platforms such as TikTok, video viewing behavior may not always represent positive feedback. Specifically, the videos are played automatically, and users passively receive the recommended videos. In this new scenario, users passively express negative feedback by skipping over videos they do not like, which provides valuable information about their preferences. Different from the negative feedback studied in traditional recommender systems, this passive-negative feedback can reflect users' interests and serve as an important supervision signal in extracting users' preferences. Therefore, it is essential to carefully design and utilize it in this novel recommendation scenario. In this work, we first conduct analyses based on a large-scale real-world short-video behavior dataset and illustrate the significance of leveraging passive feedback. We then propose a novel method that deploys the sub-interest encoder, which incorporates positive feedback and passive-negative feedback as supervision signals to learn the user's current active sub-interest. Moreover, we introduce an adaptive fusion layer to integrate various sub-interests effectively. To enhance the robustness of our model, we then introduce a multi-task learning module to simultaneously optimize two kinds of feedback - passive-negative feedback and traditional randomly-sampled negative feedback. The experiments on two large-scale datasets verify that the proposed method can significantly outperform state-of-the-art approaches. The code is released at https:// github.com/ tsinghua-fib-lab/ RecSys2023-SINE to benefit the community.	序贯推荐是推荐系统中最重要的任务之一，其目的是推荐下一个以历史行为为输入的交互项。传统的顺序推荐主要考虑收集到的积极反馈，如点击、购买等。然而，在像 TikTok 这样的短视频平台中，视频观看行为可能并不总是代表正反馈。具体来说，视频是自动播放的，用户被动地接收推荐的视频。在这个新的场景中，用户通过跳过他们不喜欢的视频被动地表达负面反馈，这提供了关于他们偏好的有价值的信息。与传统推荐系统研究的负反馈不同，这种被动负反馈能够反映用户的兴趣，是提取用户偏好的重要监督信号。因此，在这个新颖的推荐场景中仔细设计和使用它是非常重要的。在这项工作中，我们首先进行分析的基础上，大规模的现实世界的短视频行为数据集，并说明了利用被动反馈的重要性。然后我们提出了一种新的方法，部署子兴趣编码器，其中结合正反馈和被动负反馈作为监督信号，以了解用户当前的主动子兴趣。此外，我们还引入了一个自适应融合层来有效地整合各种子利益。为了提高模型的鲁棒性，我们引入了一个多任务学习模块来同时优化两种反馈-被动-负反馈和传统的随机抽样负反馈。在两个大规模数据集上的实验表明，该方法的性能明显优于目前最先进的方法。该代码在 https://github.com/tsinghua-fib-lab/RecSys2023-SINE 发布，以造福社区。	code	0
Adaptive Collaborative Filtering with Personalized Time Decay Functions for Financial Product Recommendation	Ashraf Ghiye, Baptiste Barreau, Laurent Carlier, Michalis Vazirgiannis	Ecole Polytechn, Comp Sci Lab, LIX, Palaiseau, France; BNP Paribas Corp & Inst Banking, Global Markets Data & AI Lab, Paris, France	Classical recommender systems often assume that historical data are stationary and fail to account for the dynamic nature of user preferences, limiting their ability to provide reliable recommendations in time-sensitive settings. This assumption is particularly problematic in finance, where financial products exhibit continuous changes in valuations, leading to frequent shifts in client interests. These evolving interests, summarized in the past client-product interactions, see their utility fade over time with a degree that might differ from one client to another. To address this challenge, we propose a time-dependent collaborative filtering algorithm that can adaptively discount distant client-product interactions using personalized decay functions. Our approach is designed to handle the non-stationarity of financial data and produce reliable recommendations by modeling the dynamic collaborative signals between clients and products. We evaluate our method using a proprietary dataset from BNP Paribas and demonstrate significant improvements over state-of-the-art benchmarks from relevant literature. Our findings emphasize the importance of incorporating time explicitly in the model to enhance the accuracy of financial product recommendation.	传统的推荐系统往往假设历史数据是静态的，不能解释用户偏好的动态特性，从而限制了它们在时间敏感设置中提供可靠推荐的能力。这种假设在金融领域尤其成问题，因为金融产品的估值会不断变化，导致客户利益频繁变化。这些不断发展的兴趣，总结在过去的客户-产品交互中，看到它们的效用随着时间的推移而消失，程度可能因客户而异。为了解决这个问题，我们提出了一个时间相关的协同过滤算法，可以使用个性化的衰减函数自适应地折现远距离的客户-产品相互作用。我们的方法旨在处理财务数据的非平稳性，并通过建模客户和产品之间的动态协作信号产生可靠的建议。我们使用来自法国巴黎银行的专有数据集来评估我们的方法，并证明了相对于相关文献中的最先进基准的显著改进。我们的研究结果强调了将时间明确纳入模型的重要性，以提高金融产品推荐的准确性。	code	0
Integrating Item Relevance in Training Loss for Sequential Recommender Systems	Andrea Bacciu, Federico Siciliano, Nicola Tonellotto, Fabrizio Silvestri	Univ Pisa, Pisa, Italy; Sapienza Univ Rome, Rome, Italy	Sequential Recommender Systems (SRSs) are a popular type of recommender system that leverages user history to predict the next item of interest. However, the presence of noise in user interactions, stemming from account sharing, inconsistent preferences, or accidental clicks, can significantly impact the robustness and performance of SRSs, particularly when the entire item set to be predicted is noisy. This situation is more prevalent when only one item is used to train and evaluate the SRSs. To tackle this challenge, we propose a novel approach that addresses the issue of noise in SRSs. First, we propose a sequential multi-relevant future items training objective, leveraging a loss function aware of item relevance, thereby enhancing their robustness against noise in the training data. Additionally, to mitigate the impact of noise at evaluation time, we propose multi-relevant future items evaluation (MRFI-evaluation), aiming to improve overall performance. Our relevance-aware models obtain an improvement of 1.58% of NDCG@10 and 0.96% in terms of HR@10 in the traditional evaluation protocol, the one which utilizes one relevant future item. In the MRFI-evaluation protocol, using multiple future items, the improvement is 2.82% of NDCG@10 and 0.64% of HR@10 w.r.t the best baseline model.	顺序推荐系统(SRSs)是一种流行的推荐系统，它利用用户历史来预测下一个感兴趣的项目。然而，在用户交互中存在的噪声，来自帐户共享、不一致的偏好或偶然的点击，可以显著影响 SRS 的健壮性和性能，特别是当整个项目集被预测是噪声的时候。当只有一个项目被用来训练和评估战略参考系时，这种情况更为普遍。为了应对这一挑战，我们提出了一种新的方法，解决噪音问题的 SRS。首先，我们提出了一个连续的多相关未来项目的训练目标，利用损失函数意识到项目的相关性，从而增强了他们对训练数据中的噪声的鲁棒性。此外，为了减轻噪声对评价时间的影响，我们提出了多相关的未来项目评价(MRFI- 评价) ，旨在提高整体性能。我们的相关意识模型在传统的评估方案中获得了1.58% 的 NDCG@10和0.96% 的 HR@10的改善，其中利用了一个相关的未来项目。在 MRFI 评估方案中，使用多个未来项目，改善率为 NDCG 的2.82% (10%)和 HR 的0.64% (10%)是最佳基线模型。	code	0
Integrating the ACT-R Framework with Collaborative Filtering for Explainable Sequential Music Recommendation	Marta Moscati, Christian Wallmann, Markus ReiterHaas, Dominik Kowald, Elisabeth Lex, Markus Schedl	Graz Univ Technol, Graz, Austria; Welser Profile GmbH, Gresten, Austria; Johannes Kepler Univ Linz, Inst Computat Percept, Linz, Austria	Music listening sessions often consist of sequences including repeating tracks. Modeling such relistening behavior with models of human memory has been proven effective in predicting the next track of a session. However, these models intrinsically lack the capability of recommending novel tracks that the target user has not listened to in the past. Collaborative filtering strategies, on the contrary, provide novel recommendations by leveraging past collective behaviors but are often limited in their ability to provide explanations. To narrow this gap, we propose four hybrid algorithms that integrate collaborative filtering with the cognitive architecture ACT-R. We compare their performance in terms of accuracy, novelty, diversity, and popularity bias, to baselines of different types, including pure ACT-R, kNN-based, and neural-networks-based approaches. We show that the proposed algorithms are able to achieve the best performances in terms of novelty and diversity, and simultaneously achieve a higher accuracy of recommendation with respect to pure ACT-R models. Furthermore, we illustrate how the proposed models can provide explainable recommendations.	音乐聆听课程通常包括一系列重复的曲目。利用人类记忆模型对这种重听行为进行建模已被证明对预测会话的下一个轨迹是有效的。然而，这些模型本质上缺乏推荐目标用户过去没有听过的新曲目的能力。相反，协同过滤策略通过利用过去的集体行为提供新颖的建议，但它们提供解释的能力往往有限。为了缩小这个差距，我们提出了四种混合算法，将协同过滤与认知结构 ACT-R 整合在一起。我们比较了它们在准确性、新颖性、多样性和受欢迎程度方面的表现，以及不同类型的基线，包括纯 ACT-R、基于 kNN 和基于神经网络的方法。结果表明，该算法能够在新颖性和多样性方面达到最佳性能，同时对于纯 ACT-R 模型也能够达到较高的推荐精度。此外，我们说明了所提出的模型如何能够提供可解释的建议。	code	0
An Industrial Framework for Personalized Serendipitous Recommendation in E-commerce	Zongyi Wang, Yanyan Zou, Anyu Dai, Linfang Hou, Nan Qiao, Luobao Zou, Mian Ma, Zhuoye Ding, Sulong Xu	JD com, Beijing, Peoples R China	Classical recommendation methods typically face the filter bubble problem where users likely receive recommendations of their familiar items, making them bored and dissatisfied. To alleviate such an issue, this applied paper introduces a novel framework for personalized serendipitous recommendation in an e-commerce platform (i.e., JD.com), which allows to present user unexpected and satisfying items deviating from user's prior behaviors, considering both accuracy and novelty. To achieve such a goal, it is crucial yet challenging to recognize when a user is willing to receive serendipitous items and how many novel items are expected. To address above two challenges, a two-stage framework is designed. Firstly, a DNN-based scorer is deployed to quantify the novelty degree of a product category based on user behavior history. Then, we resort to a potential outcome framework to decide the optimal timing to recommend a user serendipitous items and the novelty degree of the recommendation. Online A/B test on the e-commerce recommender platform in JD.com demonstrates that our model achieves significant gains on various metrics, 0.54% relative increase of impressive depth, 0.8% of average user click count, 3.23% and 1.38% of number of novel impressive and clicked items individually.	传统的推荐方法通常面临过滤器泡沫问题，用户可能会收到他们熟悉的项目的推荐，使他们感到厌烦和不满。为了解决这一问题，本文在电子商务平台(如京东)上引入了一个新的个性化意外推荐框架，该框架可以在考虑准确性和新颖性的情况下，提供与用户之前的行为不同的用户意想不到的、令人满意的推荐信息。要实现这样一个目标，识别用户何时愿意接收意外收获的项目以及期望接收多少新项目是至关重要的，但也是具有挑战性的。为了解决上述两个挑战，设计了一个两阶段框架。首先采用基于 DNN 的记分器，根据用户行为历史对产品类别的新颖度进行量化。然后，利用一个潜在的结果框架来决定推荐用户偶然项目的最佳时机和推荐的新颖程度。在京东的电子商务推荐平台上进行的在线 A/B 测试表明，该模型在各个指标上都取得了显著的进步，令人印象深刻的深度相对增加了0.54% ，平均用户点击次数增加了0.8% ，新颖的令人印象深刻的项目和单独点击项目的数量分别增加了3.23% 和1.38% 。	code	0
Full Index Deep Retrieval: End-to-End User and Item Structures for Cold-start and Long-tail Item Recommendation	Zhen Gong, Xin Wu, Lei Chen, Zhenzhe Zheng, Shengjie Wang, Anran Xu, Chong Wang, Fan Wu	Bytedance Inc, Mountain View, CA USA; Shanghai Jiao Tong Univ, Shanghai, Peoples R China	End-to-end retrieval models, such as Tree-based Models (TDM) and Deep Retrieval (DR), have attracted a lot of attention, but they cannot handle cold-start and long-tail item recommendation scenarios well. Specifically, DR learns a compact indexing structure, enabling efficient and accurate retrieval for large recommendation systems. However, it is discovered that DR largely fails on retrieving coldstart and long-tail items. This is because DR only utilizes user-item interaction data, which is rare and often noisy for cold-start and long-tail items. Besides, end-to-end retrieval models are unable to make use of the rich item content features. To address this issue while maintaining the efficiency of DR indexing structure, we propose Full Index Deep Retrieval (FIDR) that learns indices for the full corpus items, including cold-start and long-tail items. In addition to the original structure in DR (called User Structure in FIDR) that learns with user-item interaction data (e.g., clicks), we add an Item Structure to embed items directly based on item content features (e.g., categories). With joint efforts of User Structure and Item Structure, FIDR makes cold-start items retrievable and also improves the recommendation quality of long-tail items. To our best knowledge, FIDR is the first to solve the cold-start and longtail recommendation problem for the end-to-end retrieval models. Through extensive experiments on three real-world datasets, we demonstrate that FIDR can effectively recommend cold-start as well as long-tail items, and largely promote overall recommendation performance without sacrificing inference efficiency. According to the experiments, the recall of FIDR is improved by 8.8%similar to 11.9%, while the inference of FIDR is as efficient as DR.	端到端的检索模型，如基于树的模型(TDM)和深度检索(DR) ，已经引起了人们的广泛关注，但它们不能很好地处理冷启动和长尾项目推荐场景。具体来说，DR 学习了一种紧凑的索引结构，从而能够为大型推荐系统提供高效和准确的检索。然而，发现 DR 在检索冷启动项和长尾项时大多失败。这是因为 DR 仅利用用户项交互数据，这对于冷启动和长尾项目来说是罕见的，而且通常很吵。此外，端到端的检索模型不能利用丰富的项目内容特征。为了解决这个问题，同时保持 DR 索引结构的效率，我们提出了全索引深度检索(FIDR) ，学习完整语料库项目的索引，包括冷启动和长尾项目。除了 DR 中的原始结构(FIDR 中称为用户结构)通过用户项目交互数据(例如，点击)学习之外，我们还添加了一个项目结构来直接基于项目内容特征(例如，类别)嵌入项目。在用户结构和项目结构的共同努力下，FIDR 使冷启动项目可检索，提高了长尾项目的推荐质量。据我们所知，FIDR 首先解决了端到端检索模型的冷启动和长尾推荐问题。通过对三个实际数据集的大量实验，我们证明了 FIDR 可以有效地推荐冷启动和长尾项目，并在不牺牲推理效率的情况下大大提高整体推荐性能。实验表明，FIDR 的召回率提高了8.8% ，相当于11.9% ，而 FIDR 的推理效率与 DR 相当。	code	0
Online Matching: A Real-time Bandit System for Large-scale Recommendations	Xinyang Yi, ShaoChuan Wang, Ruining He, Hariharan Chandrasekaran, Charles Wu, Lukasz Heldt, Lichan Hong, Minmin Chen, Ed H. Chi	Google Deepmind, Mountain View, CA 94043 USA; Google Inc, Mountain View, CA USA	The last decade has witnessed many successes of deep learning-based models for industry-scale recommender systems. These models are typically trained offline in a batch manner. While being effective in capturing users' past interactions with recommendation platforms, batch learning suffers from long model-update latency and is vulnerable to system biases, making it hard to adapt to distribution shift and explore new items or user interests. Although online learning-based approaches (e.g., multi-armed bandits) have demonstrated promising theoretical results in tackling these challenges, their practical real-time implementation in large-scale recommender systems remains limited. First, the scalability of online approaches in servicing a massive online traffic while ensuring timely updates of bandit parameters poses a significant challenge. Additionally, exploring uncertainty in recommender systems can easily result in unfavorable user experience, highlighting the need for devising intricate strategies that effectively balance the trade-off between exploitation and exploration. In this paper, we introduce Online Matching: a scalable closed-loop bandit system learning from users' direct feedback on items in real time. We present a hybrid offline + online approach for constructing this system, accompanied by a comprehensive exposition of the end-to-end system architecture. We propose Diag-LinUCB - a novel extension of the LinUCB algorithm - to enable distributed updates of bandits parameter in a scalable and timely manner. We conduct live experiments in YouTube and show that Online Matching is able to enhance the capabilities of fresh content discovery and item exploration in the present platform.	过去十年见证了行业级推荐系统中基于深度学习的模型的许多成功。这些模型通常以批处理的方式离线训练。批量学习能够有效地捕捉用户过去与推荐平台的交互，但由于模型更新延迟较长，容易受到系统偏差的影响，难以适应分布变化，难以探索新的项目或用户兴趣。尽管基于在线学习的方法(例如，多武装匪徒)在应对这些挑战方面已经证明了有希望的理论成果，但它们在大规模推荐系统中的实际实时实施仍然有限。首先，在为大量在线流量提供服务的同时确保及时更新盗贼参数的在线方法的可扩展性构成了一个重大挑战。此外，在推荐系统中探索不确定性很容易导致不利的用户体验，突出需要设计复杂的策略，有效地平衡开发和勘探之间的权衡。本文介绍了在线匹配: 一个可扩展的、利用用户对项目的直接反馈进行实时学习的闭环盗窃系统。我们提出了一种混合的离线 + 在线的方法来构建这个系统，同时对端到端的系统架构进行了全面的阐述。我们提出了 Diag-LinUCB 算法—— LinUCB 算法的一个新的扩展——以便能够以可扩展和及时的方式分布式更新土匪参数。我们在 YouTube 上进行了实验，结果表明在线匹配可以增强现有平台上新内容发现和项目探索的能力。	code	0
Exploring False Hard Negative Sample in Cross-Domain Recommendation	Haokai Ma, Ruobing Xie, Lei Meng, Xin Chen, Xu Zhang, Leyu Lin, Jie Zhou	Shandong Univ, Sch Software, Jinan, Peoples R China; Tencent, WeChat, Beijing, Peoples R China	Negative Sampling in recommendation aims to capture informative negative instances for the sparse user-item interactions to improve the performance. Conventional negative sampling methods tend to select informative hard negative samples (HNS) besides the default random samples. However, these hard negative sampling methods usually struggle with false hard negative samples (FHNS), which happens when a user-item interaction has not been observed yet and is picked as a negative sample, while the user will actually interact with this item once exposed to it. Such FHNS issues may seriously confuse the model training, while most conventional hard negative sampling methods do not systematically explore and distinguish FHNS from HNS. To address this issue, we propose a novel model-agnostic Real Hard Negative Sampling (RealHNS) framework specially for cross-domain recommendation (CDR), which aims to discover the false and refine the real from all HNS via both general and cross-domain real hard negative sample selectors. For the general part, we conduct the coarse- and fine-grained real HNS selectors sequentially, armed with a dynamic item-based FHNS filter to find high-quality HNS. For the cross-domain part, we further design a new cross-domain HNS for alleviating negative transfer in CDR and discover its corresponding FHNS via a dynamic user-based FHNS filter to keep its power. We conduct experiments on four datasets based on three representative hard negative sampling methods, along with extensive model analyses, ablation studies, and universality analyses. The consistent improvements indicate the effectiveness, robustness, and universality of RealHNS, which is also easy-to-deploy in real-world systems as a plug-and-play strategy. The source code is avaliable in https://github.com/hulkima/RealHNS.	推荐中的负抽样旨在为稀疏的用户项交互捕获信息丰富的负实例，以提高性能。传统的阴性抽样方法除了选择默认随机样本外，还倾向于选择信息量大的硬阴性样本(HNS)。然而，这些硬阴性抽样方法通常与假硬阴性样本(FHNS)作斗争，这种情况发生在用户与项目的交互尚未被观察到并被选为负面样本时，而用户一旦接触到这个项目，实际上将与其交互。这样的 FHNS 问题可能会严重混淆模型训练，而大多数传统的硬阴性抽样方法没有系统地探索和区分 FHNS 和 HNS。针对这一问题，本文提出了一种针对跨域推荐(CDR)的模型无关真实硬负采样(RealHNS)框架，该框架旨在通过通用和跨域真实硬负采样选择器，发现虚假信息，并从所有 HNS 中提取真实信息。对于一般部分，我们依次进行粗粒度和细粒度的实际 HNS 选择器，并配备一个基于动态项目的 FHNS 滤波器来寻找高质量的 HNS。在跨域部分，我们进一步设计了一个新的跨域 HNS 来减轻 CDR 中的负转移，并通过一个基于用户的动态 FHNS 滤波器来发现相应的 FHNS 以保持其功率。基于三种典型的硬负取样方法，我们对四个数据集进行了实验，同时进行了广泛的模型分析、烧蚀研究和通用性分析。一致的改进表明了 RealHNS 的有效性、健壮性和通用性，作为一种即插即用策略，它也很容易在现实世界的系统中部署。源代码有 https://github.com/hulkima/realhns。	code	0
Contrastive Learning with Frequency-Domain Interest Trends for Sequential Recommendation	Yichi Zhang, Guisheng Yin, Yuxin Dong	Harbin Engn Univ, Harbin, Heilongjiang, Peoples R China	Recently, contrastive learning for sequential recommendation has demonstrated its powerful ability to learn high-quality user representations. However, constructing augmented samples in the time domain poses challenges due to various reasons, such as fast-evolving trends, interest shifts, and system factors. Furthermore, the F-principle indicates that deep learning preferentially fits the low-frequency part, resulting in poor performance on high-frequency tasks. The complexity of time series and the low-frequency preference limit the utility of sequence encoders. To address these challenges, we need to construct augmented samples from the frequency domain, thus improving the ability to accommodate events of different frequency sizes. To this end, we propose a novel Contrastive Learning with Frequency-Domain Interest Trends for Sequential Recommendation (CFIT4SRec). We treat the embedding representations of historical interactions as "images" and introduce the secondorder Fourier transform to construct augmented samples. The components of different frequency sizes reflect the interest trends between attributes and their surroundings in the hidden space. We introduce three data augmentation operations to accommodate events of different frequency sizes: low-pass augmentation, high-pass augmentation, and band-stop augmentation. Extensive experiments on four public benchmark datasets demonstrate the superiority of CFIT4SRec over the state-of-the-art baselines. The implementation code is available at https://github.com/zhangyichi1Z/CFIT4SRec.	近年来，序贯推荐的对比学习已经证明了其学习高质量用户表示的强大能力。然而，由于各种原因，如快速发展的趋势、兴趣转移和系统因素等，在时间域构造增广样本提出了挑战。此外，F- 原理表明，深度学习优先适用于低频部分，导致高频任务的性能较差。时间序列的复杂性和低频偏好限制了序列编码器的实用性。为了应对这些挑战，我们需要从频率域构造增强样本，从而提高容纳不同频率大小事件的能力。为此，我们提出了一种新的对比学习与频域兴趣趋势顺序推荐(CFIT4SRec)。我们把历史相互作用的嵌入表示当作“图像”，并引入二阶傅里叶变换来构造增广样本。不同频率大小的分量反映了隐藏空间中属性与环境之间的兴趣趋势。我们引入三种数据增强操作来适应不同频率大小的事件: 低通增强、高通增强和带阻增强。在四个公共基准数据集上的大量实验证明了 CFIT4SRec 相对于最先进的基准线的优越性。实施守则可于 https://github.com/zhangyichi1z/cfit4srec 索取。	code	0
Multi-task Item-attribute Graph Pre-training for Strict Cold-start Item Recommendation	Yuwei Cao, Liangwei Yang, Chen Wang, Zhiwei Liu, Hao Peng, Chenyu You, Philip S. Yu	Yale Univ, New Haven, CT USA; Univ Illinois, Chicago, IL USA; Salesforce AI, Washington, DC USA; Beihang Univ, Beijing, Peoples R China	Recommendation systems suffer in the strict cold-start (SCS) scenario, where the user-item interactions are entirely unavailable. The well-established, dominating identity (ID)-based approaches completely fail to work. Cold-start recommenders, on the other hand, leverage item contents ( brand, title, descriptions, etc.) to map the new items to the existing ones. However, the existing SCS recommenders explore item contents in coarse-grained manners that introduce noise or information loss. Moreover, informative data sources other than item contents, such as users' purchase sequences and review texts, are largely ignored. In this work, we explore the role of the fine-grained item attributes in bridging the gaps between the existing and the SCS items and pre-train a knowledgeable item-attribute graph for SCS item recommendation. Our proposed framework, ColdGPT, models item-attribute correlations into an item-attribute graph by extracting fine-grained attributes from item contents. ColdGPT then transfers knowledge into the item-attribute graph from various available data sources, i.e., item contents, historical purchase sequences, and review texts of the existing items, via multi-task learning. To facilitate the positive transfer, ColdGPT designs specific submodules according to the natural forms of the data sources and proposes to coordinate the multiple pre-training tasks via unified alignment-and-uniformity losses. Our pre-trained item-attribute graph acts as an implicit, extendable item embedding matrix, which enables the SCS item embeddings to be easily acquired by inserting these items into the item-attribute graph and propagating their attributes' embeddings. We carefully process three public datasets, i.e., Yelp, Amazon-home, and Amazon-sports, to guarantee the SCS setting for evaluation. Extensive experiments show that ColdGPT consistently outperforms the existing SCS recommenders by large margins and even surpasses models that are pre-trained on 75 - 224 times more, cross-domain data on two out of four datasets. Our code and pre-processed datasets for SCS evaluations are publicly available to help future SCS studies.	推荐系统在严格的冷启动(SCS)场景中受到影响，其中用户-项交互是完全不可用的。基于身份(ID)的成熟的、占主导地位的方法完全不起作用。另一方面，冷启动推荐器利用项目内容(品牌、标题、描述等)将新项目映射到现有项目。然而，现有的 SCS 推荐标准以粗粒度的方式探索项目内容，导致噪声或信息丢失。此外，除了项目内容之外的信息性数据源，如用户的购买顺序和评论文本，在很大程度上被忽略。在本研究中，我们探讨细粒度项目属性在弥补现有项目与 SCS 项目之间的差距方面所起的作用，并预先训练出一个知识化的项目属性图来进行 SCS 项目推荐。我们提出的框架 ColdGPT 通过从项目内容中提取细粒度属性，将项目-属性关系建模成项目-属性图。ColdGPT 然后通过多任务学习，将来自各种可用数据源的知识转移到项目属性图中，即项目内容、历史购买顺序和审查现有项目的文本。为了便于正向传输，ColdGPT 根据数据源的自然形式设计了具体的子模块，并提出通过统一的对齐和一致性损失来协调多个预训练任务。我们预先训练的项目属性图作为一个隐式的、可扩展的项目嵌入矩阵，通过将这些项目插入到项目属性图中并传播它们的属性嵌入，可以方便地获得 SCS 项目嵌入。我们仔细处理三个公共数据集，即 Yelp、 Amazon-home 和 Amazon-sports，以保证 SCS 设置用于评估。大量的实验表明，ColdGPT 始终优于现有的 SCS 推荐器，甚至超过预先训练75-224倍以上的模型，跨域数据在四个数据集中的两个。我们的代码和 SCS 评估的预处理数据集是公开的，以帮助未来的 SCS 研究。	code	0
BVAE: Behavior-aware Variational Autoencoder for Multi-Behavior Multi-Task Recommendation	Qianzhen Rao, Yang Liu, Weike Pan, Zhong Ming	Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen, Peoples R China	A practical recommender system should be able to handle heterogeneous behavioral feedback as inputs and has multi-task outputs ability. Although the heterogeneous one-class collaborative filtering (HOCCF) and multi-task learning (MTL) methods has been well studied, there is still a lack of targeted manner in their combined fields, i.e., Multi-behavior Multi-task Recommendation (MMR). To fill the gap, we propose a novel recommendation framework called Behavior-aware Variational AutoEncoder (BVAE), which meliorates the parameter sharing and loss minimization method with the VAE structure to address the MMR problem. Specifically, our BVAE includes behavior-aware semi-encoders and decoders, and a target feature fusion network with a global feature filtering network, while using standard deviation to weigh loss. These modules generate the behavior-aware recommended item list via constructing better semantic feature vectors for users, i.e., from dual perspectives of behavioral preference and global interaction. In addition, we optimize our BVAE in terms of adaptability and robustness, i.e., it is concise and flexible to consume any amount of behaviors with different distributions. Extensive empirical studies on two real and widely used datasets confirm the validity of our design and show that our BVAE can outperform the state-of-the-art related baseline methods under multiple evaluation metrics. The processed datasets, source code, and scripts necessary to reproduce the results can be available at https://github.com/WitnessForest/BVAE.	一个实际的推荐系统应该能够处理异质的行为反馈作为输入，并具有多任务输出的能力。虽然单一类别协同过滤(HOCCF)和多任务学习(MTL)方法已经得到了很好的研究，但是在它们的组合领域，即多行为多任务推荐(mMR) ，仍然缺乏有针对性的方式。为了填补这一空白，我们提出了一种新的推荐框架，称为行为感知变量自动编码器(BVAE) ，它改进了参数共享和损失最小化方法与 VAE 结构，以解决 MMR 问题。具体来说，我们的 BVAE 包括行为感知半编码器和解码器，目标特征融合网络与全球特征过滤网络，同时使用标准差来衡量损失。这些模块通过为用户构造更好的语义特征向量，即从行为偏好和全局交互的双重视角生成行为感知的推荐项目列表。此外，我们在适应性和健壮性方面对 BVAE 进行了优化，也就是说，使用不同分布的任何数量的行为都是简洁和灵活的。对两个实际和广泛使用的数据集的大量实证研究证实了我们的设计的有效性，并表明我们的 BVAE 能够在多个评估指标下超越最先进的相关基线方法。处理过的数据集、源代码和重现结果所需的脚本可以在 https://github.com/witnessforest/bvae 上找到。	code	0
Looks Can Be Deceiving: Linking User-Item Interactions and User's Propensity Towards Multi-Objective Recommendations	Patrik Dokoupil, Ladislav Peska, Ludovico Boratto	Charles Univ Prague, Fac Math & Phys, Prague, Czech Republic; Univ Cagliari, Cagliari, Italy	Multi-objective recommender systems (MORS) provide suggestions to users according to multiple (and possibly conflicting) goals. When a system optimizes its results at the individual-user level, it tailors them on a user's propensity towards the different objectives. Hence, the capability to understand users' fine-grained needs towards each goal is crucial. In this paper, we present the results of a user study in which we monitored the way users interacted with recommended items, as well as their self-proclaimed propensities towards relevance, novelty, and diversity objectives. The study was divided into several sessions, where users evaluated recommendation lists originating from a relevance-only single-objective baseline as well as MORS. We show that, despite MORS-based recommendations attracting fewer selections, their presence in the early sessions are crucial for users' satisfaction in the later stages. Surprisingly, the self-proclaimed willingness of users to interact with novel and diverse items is not always reflected in the recommendations they accept. Post-study questionnaires provide insights on how to deal with this matter, suggesting that MORS-based results should be accompanied by elements that allow users to understand the recommendations, so as to facilitate the choice of whether a recommendation should be accepted or not. Detailed study results are available at https://bit.ly/looks-can-be-deceiving-repo.	多目标推荐系统(MORS)根据多个(可能存在冲突的)目标向用户提供建议。当一个系统在个人用户层面优化其结果时，它会根据用户对不同目标的倾向来调整结果。因此，理解用户对每个目标的细粒度需求的能力是至关重要的。在本文中，我们介绍了一项用户研究的结果，其中我们监测用户与推荐项目的互动方式，以及他们自称的相关性，新颖性和多样性的目标的倾向。这项研究分为几个阶段，用户评估源自相关性单一目标基线的推荐名单以及 MORS。我们表明，尽管基于 MORS 的推荐吸引了较少的选择，但是它们在早期会话中的出现对于用户在后期阶段的满意度是至关重要的。令人惊讶的是，用户自称愿意与新颖和多样化的项目进行互动，但并不总是反映在他们接受的建议中。研究后调查问卷提供了关于如何处理这一问题的见解，表明基于 MORS 的结果应该伴随着使用者能够理解建议的要素，以便于选择是否接受建议。详细研究结果载于 https://bit.ly/looks-can-be-deceiving-repo。	code	0
Scaling Session-Based Transformer Recommendations using Optimized Negative Sampling and Loss Functions	Timo Wilm, Philipp Normann, Sophie Baumeister, PaulVincent Kobow	OTTO GmbH & Co KG, Hamburg, Germany	This work introduces TRON, a scalable session-based Transformer Recommender using Optimized Negative-sampling. Motivated by the scalability and performance limitations of prevailing models such as SASRec and GRU4Rec(+), TRON integrates top-k negative sampling and listwise loss functions to enhance its recommendation accuracy. Evaluations on relevant large-scale e-commerce datasets show that TRON improves upon the recommendation quality of current methods while maintaining training speeds similar to SAS-Rec. A live A/B test yielded an 18.14% increase in click-through rate over SASRec, highlighting the potential of TRON in practical settings. For further research, we provide access to our source code(1) and an anonymized dataset(2).	本文介绍了 TRON，一个可扩展的基于会话的变压器优化负采样推荐器。由于 SASRec 和 GRU4Rec (+)等主流模型的可扩展性和性能限制，TRON 集成了 top-k 负采样和列表损失功能，以提高其推荐的准确性。对相关大规模电子商务数据集的评估表明，TRON 在保持与 SAS-Rec 类似的训练速度的同时，提高了现有方法的推荐质量。现场 A/B 测试的点进率比 SASrec 增加了18.14% ，突出了 TRON 在实际环境中的潜力。为了进一步研究，我们提供了对源代码(1)和匿名数据集(2)的访问。	code	0
Pairwise Intent Graph Embedding Learning for Context-Aware Recommendation	Dugang Liu, Yuhao Wu, Weixin Li, Xiaolian Zhang, Hao Wang, Qinjuan Yang, Zhong Ming	Huawei 2012 Lab, Shenzhen, Peoples R China; Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen, Peoples R China; Shenzhen Univ, Guangdong Lab Artificial Intelligence & Digital E, Shenzhen, Peoples R China	Although knowledge graph has shown their effectiveness in mitigating data sparsity in many recommendation tasks, they remain underutilized in context-aware recommender systems (CARS) with the specific sparsity challenges associated with the contextual features, i.e., feature sparsity and interaction sparsity. To bridge this gap, in this paper, we propose a novel pairwise intent graph embedding learning (PING) framework to efficiently integrate knowledge graphs into CARS. Specifically, our PING contains three modules: 1) a graph construction module is used to obtain a pairwise intent graph (PIG) containing nodes for users, items, entities, and enhanced intent, where enhanced intent nodes are generated by applying user intent fusion (UIF) on relational intent and contextual intent, and two sub-intents are derived from the semantic information and contextual information, respectively; 2) a pairwise intent joint graph convolution module is used to obtain the refined embeddings of all the features by executing a customized convolution strategy on PIG, where each enhanced intent node acts as a hub to efficiently propagate information among different features and between all the features and knowledge graph; 3) a recommendation module with the refined embeddings is used to replace the randomly initialized embeddings of downstream recommendation models to improve model performance. Finally, we conduct extensive experiments on three public datasets to verify the effectiveness and compatibility of our PING.	尽管知识图表显示了它们在许多推荐任务中缓解数据稀疏的有效性，但是它们在上下文感知的推荐系统(CARS)中仍然没有得到充分利用，与上下文特征相关的特定稀疏性挑战，即特征稀疏和交互稀疏。为了弥补这一差距，本文提出了一种新的成对意图嵌入学习(PING)框架，有效地将知识图集成到 CARS 中。具体来说，我们的 pING 包含三个模块: 1)一个图形构造模块用于获得包含用户、项目、实体和增强意图节点的成对意图图(pIG) ，其中增强意图节点是通过在关系意图和上下文意图上应用用户意图融合(UIF)而生成的，并且两个子意图分别来自语义信息和上下文信息;2)通过在 PIG 上执行定制的卷积策略，使用成对意图联合图卷积模块来获得所有特征的精细嵌入，其中每个增强意图节点作为一个中心，在不同特征之间以及所有特征和知识图之间有效地传播信息; 3)使用具有精细嵌入的推荐模块来代替下游推荐模型的随机初始化嵌入，以提高模型性能。最后，我们在三个公共数据集上进行了广泛的实验，以验证我们的 PING 的有效性和兼容性。	code	0
A Model-Agnostic Framework for Recommendation via Interest-aware Item Embeddings	Amit Kumar Jaiswal, Yu Xiong		Item representation holds significant importance in recommendation systems, which encompasses domains such as news, retail, and videos. Retrieval and ranking models utilise item representation to capture the user-item relationship based on user behaviours. While existing representation learning methods primarily focus on optimising item-based mechanisms, such as attention and sequential modelling. However, these methods lack a modelling mechanism to directly reflect user interests within the learned item representations. Consequently, these methods may be less effective in capturing user interests indirectly. To address this challenge, we propose a novel Interest-aware Capsule network (IaCN) recommendation model, a model-agnostic framework that directly learns interest-oriented item representations. IaCN serves as an auxiliary task, enabling the joint learning of both item-based and interest-based representations. This framework adopts existing recommendation models without requiring substantial redesign. We evaluate the proposed approach on benchmark datasets, exploring various scenarios involving different deep neural networks, behaviour sequence lengths, and joint learning ratios of interest-oriented item representations. Experimental results demonstrate significant performance enhancements across diverse recommendation models, validating the effectiveness of our approach.	项目表示在推荐系统中具有重要意义，推荐系统包括新闻、零售和视频等领域。检索和排序模型利用项目表示来捕获基于用户行为的用户-项目关系。而现有的表征学习方法主要集中在优化项目为基础的机制，如注意力和顺序建模。然而，这些方法缺乏直接反映用户兴趣的建模机制。因此，这些方法在间接捕获用户兴趣方面可能不太有效。为了应对这一挑战，我们提出了一种新的兴趣感知胶囊网络(IaCN)推荐模型，这是一个直接学习兴趣导向的项目表示的模型无关框架。IaCN 作为辅助任务，支持基于项目和基于兴趣的表示的联合学习。该框架采用现有的推荐模型，无需重新设计。我们评估了所提出的基准数据集方法，探索了涉及不同深度神经网络、行为序列长度和兴趣导向项目表示的联合学习比率的各种场景。实验结果显示了不同推荐模型的性能显著提高，验证了我们方法的有效性。	code	0
Gradient Matching for Categorical Data Distillation in CTR Prediction	Cheng Wang, Jiacheng Sun, Zhenhua Dong, Ruixuan Li, Rui Zhang	Ruizhang Info, Shenzhen, Peoples R China; Huazhong Univ Sci & Technol, Wuhan, Peoples R China; Huawei Noahs Ark Lab, Shenzhen, Peoples R China	The cost of hardware and energy consumption on training a click-through rate (CTR) model is highly prohibitive. A recent promising direction for reducing such costs is data distillation with gradient matching, which aims to synthesize a small distilled dataset to guide the model to a similar parameter space as those trained on real data. However, there are two main challenges to implementing such a method in the recommendation field: (1) The categorical recommended data are high dimensional and sparse one- or multi-hot data which will block the gradient flow, causing backpropagation-based data distillation invalid. (2) The data distillation process with gradient matching is computationally expensive due to the bi-level optimization. To this end, we investigate efficient data distillation tailored for recommendation data with plenty of side information where we formulate the discrete data to the dense and continuous data format. Then, we further introduce a one-step gradient matching scheme, which performs gradient matching for only a single step to overcome the inefficient training process. The overall proposed method is called Categorical data distillation with Gradient Matching (CGM), which is capable of distilling a large dataset into a small of informative synthetic data for training CTR models from scratch. Experimental results show that our proposed method not only outperforms the state-of-the-art coreset selection and data distillation methods but also has remarkable cross-architecture performance. Moreover, we explore the application of CGM on model retraining and mitigate the effect of different random seeds on the training results.	培训点进率模型的硬件和能源消耗成本高得令人望而却步。最近一个有希望降低这种成本的方向是使用梯度匹配的数据提取，其目的是合成一个小的提取的数据集，以引导模型到一个类似的参数空间，因为这些训练的实际数据。然而，这种方法在推荐领域的实现面临两个主要挑战: (1)分类推荐数据是高维稀疏的一个或多个热点数据，会阻塞梯度流，导致基于反向传播的数据精馏失效。(2)采用梯度匹配的数据精馏过程，由于采用了双层优化算法，计算量较大。为此，我们研究了针对推荐数据的有效数据精馏，这些推荐数据具有丰富的侧信息，我们将离散数据表述为密集和连续的数据格式。然后，我们进一步引入了一个一步梯度匹配方案，该方案仅对一个步骤进行梯度匹配，以克服训练过程的低效性。提出了一种基于梯度匹配(CGM)的分类数据提取方法，该方法能够将大量的数据集提取为一小部分信息量大的综合数据，用于从头开始训练 CTR 模型。实验结果表明，该方法不仅优于现有的复位选择和数据提取方法，而且具有显著的交叉结构性能。此外，我们还探讨了 CGM 在模型再训练中的应用，以减轻不同随机种子对训练结果的影响。	code	0
Augmented Negative Sampling for Collaborative Filtering	Yuhan Zhao, Rui Chen, Riwei Lai, Qilong Han, Hongtao Song, Li Chen	Hong Kong Baptist Univ, Hong Kong, Peoples R China; Harbin Engn Univ, Harbin, Peoples R China	Negative sampling is essential for implicit-feedback-based collaborative filtering, which is used to constitute negative signals from massive unlabeled data to guide supervised learning. The state-of-the-art idea is to utilize hard negative samples that carry more useful information to form a better decision boundary. To balance efficiency and effectiveness, the vast majority of existing methods follow the two-pass approach, in which the first pass samples a fixed number of unobserved items by a simple static distribution and then the second pass selects the final negative items using a more sophisticated negative sampling strategy. However, selecting negative samples from the original items in a dataset is inherently restricted due to the limited available choices, and thus may not be able to contrast positive samples well. In this paper, we confirm this observation via carefully designed experiments and introduce two major limitations of existing solutions: ambiguous trap and information discrimination. Our response to such limitations is to introduce "augmented" negative samples that may not exist in the original dataset. This direction renders a substantial technical challenge because constructing unconstrained negative samples may introduce excessive noise that eventually distorts the decision boundary. To this end, we introduce a novel generic augmented negative sampling (ANS) paradigm and provide a concrete instantiation. First, we disentangle hard and easy factors of negative items. Next, we generate new candidate negative samples by augmenting only the easy factors in a regulated manner: the direction and magnitude of the augmentation are carefully calibrated. Finally, we design an advanced negative sampling strategy to identify the final augmented negative samples, which considers not only the score function used in existing methods but also a new metric called augmentation gain. Extensive experiments on real-world datasets demonstrate that our method significantly outperforms state-of-the-art baselines. Our code is publicly available at https://github.com/Asa9aoTK/ANS-Recbole.	对于基于内隐反馈的协同过滤来说，负采样是必不可少的，它用来从大量未标记的数据中构成负信号来引导监督式学习。最先进的想法是利用带有更多有用信息的硬阴性样本来形成更好的决策边界。为了平衡效率和有效性，绝大多数现有的方法遵循双通道方法，其中第一通道通过一个简单的静态分布采样一个固定数量的未观测项目，然后第二通道选择最终的负项目使用一个更复杂的负面采样策略。然而，从数据集中的原始项目中选择阴性样本本质上受到限制，因为可用的选择有限，因此可能无法很好地对比阳性样本。在本文中，我们通过精心设计的实验证实了这一观察，并介绍了现有解决方案的两个主要局限性: 模糊陷阱和信息辨别。我们对这些限制的反应是引入“增强”的负样本，这些样本可能不存在于原始数据集中。这种方向带来了巨大的技术挑战，因为构建无约束的负样本可能会引入过多的噪音，最终导致决策边界失真。为此，我们引入了一个新的通用增广负抽样(ANS)范式，并提供了一个具体的实例。首先，我们对消极项目中的难易因素进行了解析。接下来，我们通过以一种有规律的方式增加简单因子来产生新的候选阴性样本: 增加的方向和幅度被仔细校准。最后，我们设计了一个先进的负抽样策略来识别最终的增广负样本，它不仅考虑了现有方法中使用的得分函数，而且还考虑了一个新的度量称为增广增益。对真实世界数据集的大量实验表明，我们的方法明显优于最先进的基线。我们的代码可以在 https://github.com/asa9aotk/ans-recbole 上公开获取。	code	0
LightSAGE: Graph Neural Networks for Large Scale Item Retrieval in Shopee's Advertisement Recommendation	Dang Minh Nguyen, Chenfei Wang, Yan Shen, Yifan Zeng	SEA Grp, Shopee, Beijing, Peoples R China; SEA Grp, Shopee, Singapore, Singapore	Graph Neural Network (GNN) is the trending solution for item retrieval in recommendation problems. Most recent reports, however, focus heavily on new model architectures. This may bring some gaps when applying GNN in the industrial setup, where, besides the model, constructing the graph and handling data sparsity also play critical roles in the overall success of the project. In this work, we report how GNN is applied for large-scale e-commerce item retrieval at Shopee. We introduce our simple yet novel and impactful techniques in graph construction, modeling, and handling data skewness. Specifically, we construct high-quality item graphs by combining strong-signal user behaviors with high-precision collaborative filtering (CF) algorithm. We then develop a new GNN architecture named LightSAGE to produce high-quality items' embeddings for vector search. Finally, we design multiple strategies to handle cold-start and long-tail items, which are critical in an advertisement (ads) system. Our models bring improvement in offline evaluations, online A/B tests, and are deployed to the main traffic of Shopee's Recommendation Advertisement system.	图形神经网络(GNN)是推荐问题中项目检索的趋势解决方案。然而，最近的大多数报告主要关注于新的模型架构。这可能会带来一些差距时，GNN 在工业设置，其中，除了模型，构造图和处理数据稀疏也发挥关键作用的项目的整体成功。在这项工作中，我们报告了 GNN 是如何应用于 Shopee 的大规模电子商务项目检索。我们介绍了我们在图形构造、建模和处理数据偏态方面的简单而新颖且有影响力的技术。具体来说，我们通过结合强信号用户行为和高精度协同过滤(CF)算法来构建高质量的项目图。然后，我们开发了一个名为 LightSAGE 的新 GNN 体系结构，以生成用于向量搜索的高质量条目嵌入。最后，我们设计了多种策略来处理冷启动和长尾项目，这是一个广告系统的关键。我们的模型带来了线下评估和在线 A/B 测试的改进，并被部署到 Shopee 推荐广告系统的主要流量中。	code	0
Goal-Oriented Multi-Modal Interactive Recommendation with Verbal and Non-Verbal Relevance Feedback	Yaxiong Wu, Craig Macdonald, Iadh Ounis	Univ Glasgow, Glasgow, Lanark, Scotland	Interactive recommendation enables users to provide verbal and non-verbal relevance feedback (such as natural-language critiques and likes/dislikes) when viewing a ranked list of recommendations (such as images of fashion products), in order to guide the recommender system towards their desired items (i.e. goals) across multiple interaction turns. Such a multi-modal interactive recommendation (MMIR) task has been successfully formulated with deep reinforcement learning (DRL) algorithms by simulating the interactions between an environment (i.e. a user) and an agent (i.e. a recommender system). However, it is typically challenging and unstable to optimise the agent to improve the recommendation quality associated with implicit learning of multi-modal representations in an end-to-end fashion in DRL. This is known as the coupling of policy optimisation and representation learning. To address this coupling issue, we propose a novel goal-oriented multi-modal interactive recommendation model (GOMMIR) that uses both verbal and non-verbal relevance feedback to effectively incorporate the users' preferences over time. Specifically, our GOMMIR model employs a multi-task learning approach to explicitly learn the multi-modal representations using a multi-modal composition network when optimising the recommendation agent. Moreover, we formulate the MMIR task using goal-oriented reinforcement learning and enhance the optimisation objective by leveraging non-verbal relevance feedback for hard negative sampling and providing extra goal-oriented rewards to effectively optimise the recommendation agent. Following previous work, we train and evaluate our GOMMIR model by using user simulators that can generate natural-language feedback about the recommendations as a surrogate for real human users. Experiments conducted on four well-known fashion datasets demonstrate that our proposed GOMMIR model yields significant improvements in comparison to the existing state-of-the-art baseline models.	交互式推荐可以让用户在浏览排名推荐列表(如时尚产品图片)时提供语言和非语言关联反馈(如自然语言评论和喜欢/不喜欢) ，以便引导推荐系统在多个互动回合中朝着他们想要的项目(即目标)前进。这样一个多模态交互推荐(MMIR)任务已经成功地通过深度强化学习(DRL)算法模拟了环境(比如用户)和代理(比如推荐系统)之间的交互。然而，在 DRL 中以端到端的方式优化代理以提高与多模态表示的隐式学习相关的推荐质量通常具有挑战性和不稳定性。这就是所谓的政策优化和表示学习的耦合。为了解决这个耦合问题，我们提出了一个新的目标导向的多模式交互式推荐模型(GOMMIR) ，它使用语言和非语言的关联反馈来有效地整合用户的喜好随着时间的推移。我们的 GOMMIR 模型采用多任务学习方法，在优化推荐代理时使用多模态组合网络显式学习多模态表示。此外，我们利用目标导向的强化学习制定 MMIR 任务，并利用非语言关联反馈进行硬性负面抽样，以及提供额外的目标导向奖励，以提高优化目标，从而有效地优化推荐代理。在以前的工作之后，我们训练和评估我们的 GOMMIR 模型，使用用户模拟器，可以生成自然语言反馈的建议作为一个替代真正的人类用户。在四个著名的时尚数据集上进行的实验表明，我们提出的 GOMMIR 模型与现有的最先进的基线模型相比，产生了显著的改进。	code	0
DREAM: Decoupled Representation via Extraction Attention Module and Supervised Contrastive Learning for Cross-Domain Sequential Recommender	Xiaoxin Ye, Yun Li, Lina Yao	CSIROs Data61, Sydney, NSW, Australia; UNSW, Sch Comp Sci & Engn, Sydney, NSW, Australia	Cross-Domain Sequential Recommendation(CDSR) aims to generate accurate predictions for future interactions by leveraging users' cross-domain historical interactions. One major challenge of CDSR is howto jointly learn the single- and cross-domain user preferences efficiently. To enhance the target domain's performance, most existing solutions start by learning the single-domain user preferences within each domain and then transferring the acquired knowledge from the rich domain to the target domain. However, this approach ignores the inter-sequence item relationship and also limits the opportunities for target domain knowledge to enhance the rich domain performance. Moreover, it also ignores the information within the cross-domain sequence. Despite cross-domain sequences being generally noisy and hard to learn directly, they contain valuable user behavior patterns with great potential to enhance performance. Another key challenge of CDSR is data sparsity, which also exists in other recommendation system problems. In the real world, the data distribution of the recommendation system is highly skewed to the popular products, especially on the large-scale dataset with millions of users and items. One more challenge is the class imbalance problem, inherited by the sequential recommendation problem. Generally, each sample only has one positive and thousands of negative samples. To address the above problems together, an innovative Decoupled Representation via Extraction Attention Module (DREAM) is proposed for CDSR to simultaneously learn singleand cross-domain user preference via decoupled representations. A novel Supervised Contrastive Learning framework is introduced to model the inter-sequence relationship as well as address the data sparsity via data augmentations. DREAM also leverages Focal Loss to put more weight on misclassified samples to address the class-imbalance problem, with another uplift on the overall model performance. Extensive experiments had been conducted on two cross-domain recommendation datasets, demonstrating DREAM outperforms various SOTA cross-domain recommendation algorithms achieving up to a 75% uplift in Movie-Book Scenarios.	跨域序列推荐(CDSR)旨在通过利用用户的跨域历史交互产生对未来交互的准确预测。CDSR 的一个主要挑战是如何有效地联合学习单域和跨域用户偏好。为了提高目标领域的性能，大多数现有的解决方案都是从学习每个领域内的单领域用户偏好开始，然后将获得的知识从富领域转移到目标领域。然而，这种方法忽略了序列间的项目关系，同时也限制了目标领域知识提高丰富领域性能的机会。此外，它还忽略了跨域序列中的信息。尽管跨域序列通常有噪声且难以直接学习，但它们包含有价值的用户行为模式，具有提高性能的巨大潜力。CDSR 的另一个关键挑战是数据稀疏性，这也存在于其他推荐系统问题中。在现实世界中，推荐系统的数据分布高度偏向于流行产品，特别是在拥有数百万用户和项目的大规模数据集上。另一个挑战是由顺序推荐问题继承的类不平衡问题。一般来说，每个样本只有一个阳性和数千个阴性样本。针对上述问题，提出了一种新的基于抽取注意模块的解耦表示方法(DREAM) ，使 CDSR 能够通过解耦表示同时学习单域和跨域用户偏好。提出了一种新的有监督对比学习框架，通过数据增强对序列间的关系进行建模，并解决了数据稀疏问题。梦想也利用焦损更加重视错误分类的样本，以解决类不平衡的问题，与另一个提升整体模型的性能。在两个跨域推荐数据集上进行了广泛的实验，证明了 DREAM 优于各种 SOTA 跨域推荐算法，在电影图书场景中实现了高达75% 的提升。	code	0
A Multi-view Graph Contrastive Learning Framework for Cross-Domain Sequential Recommendation	Zitao Xu, Weike Pan, Zhong Ming	Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen, Peoples R China	Sequential recommendation methods play an irreplaceable role in recommender systems which can capture the users' dynamic preferences from the behavior sequences. Despite their success, these works usually suffer from the sparsity problem commonly existed in real applications. Cross-domain sequential recommendation aims to alleviate this problem by introducing relatively richer source-domain data. However, most existing methods capture the users' preferences independently of each domain, which may neglect the item transition patterns across sequences from different domains, i.e., a user's interaction in one domain may influence his/her next interaction in other domains. Moreover, the data sparsity problem still exists since some items in the target and source domains are interacted with only a limited number of times. To address these issues, in this paper we propose a generic framework named multi-view graph contrastive learning (MGCL). Specifically, we adopt the contrastive mechanism in an intra-domain item representation view and an inter-domain user preference view. The former is to jointly learn the dynamic sequential information in the user sequence graph and the static collaborative information in the cross-domain global graph, while the latter is to capture the complementary information of the user's preferences from different domains. Extensive empirical studies on three real-world datasets demonstrate that our MGCL significantly outperforms the state-of-the-art methods.	序列推荐方法在推荐系统中起着不可替代的作用，它可以从用户的行为序列中获取用户的动态偏好。尽管这些作品取得了成功，但它们在实际应用中普遍存在着稀疏性问题。跨域顺序推荐旨在通过引入相对丰富的源域数据来缓解这一问题。然而，大多数现有的方法捕获用户的偏好独立于每个领域，这可能会忽略来自不同领域的序列之间的项目转换模式，也就是说，一个用户在一个领域的交互可能会影响他/她在其他领域的下一个交互。此外，由于目标域和源域中的某些项目只能进行有限次数的交互，因此数据稀疏问题仍然存在。为了解决这些问题，本文提出了一种通用的多视图图形对比学习(MGCL)框架。具体来说，我们在域内项表示视图和域间用户首选项视图中采用了对比机制。前者是联合学习用户序列图中的动态序列信息和跨域全局图中的静态协作信息，后者是从不同领域获取用户偏好的互补信息。对三个真实世界数据集的大量实证研究表明，我们的 MGCL 明显优于最先进的方法。	code	0
STAN: Stage-Adaptive Network for Multi-Task Recommendation by Learning User Lifecycle-Based Representation	Wanda Li, Wenhao Zheng, Xuanji Xiao, Suhang Wang	Penn State Univ, University Pk, PA 16802 USA; Shopee Co, Beijing, Peoples R China; Tsinghua Univ, Beijing, Peoples R China	Recommendation systems play a vital role in many online platforms, with their primary objective being to satisfy and retain users. As directly optimizing user retention is challenging, multiple evaluation metrics are often employed. Current methods often use multi-task learning to optimize these measures. However, they usually miss that users have personal preferences for different tasks, which can change over time. Identifying and tracking the evolution of user preferences can lead to better user retention. To address this issue, we introduce the concept of "user lifecycle," consisting of multiple stages characterized by users' varying preferences for different tasks. We propose a novel Stage-Adaptive Network (STAN) framework for modeling user lifecycle stages. STAN first identifies latent user lifecycle stages based on learned user preferences and then employs the stage representation to enhance multi-task learning performance. Our experimental results using both public and industrial datasets demonstrate that the proposed model significantly improves multi-task prediction performance compared to state-of-the-art methods, highlighting the importance of considering user lifecycle stages in recommendation systems. Online A/B testing reveals that our model outperforms the existing model, achieving a significant improvement of 3.05% in staytime per user and 0.88% in CVR. We have deployed STAN on all Shopee live-streaming recommendation services.	推荐系统在许多在线平台中发挥着至关重要的作用，其主要目标是满足和留住用户。由于直接优化用户保留是具有挑战性的，因此经常采用多种评估指标。目前的方法往往采用多任务学习来优化这些措施。然而，他们通常忽略了用户对不同任务的个人偏好，这种偏好可能随着时间的推移而改变。识别和跟踪用户偏好的演变可以更好地保留用户。为了解决这个问题，我们引入了“用户生命周期”的概念，包括多个阶段，拥有属性用户对不同任务的不同偏好。我们提出了一个新的阶段自适应网络(STAN)框架，用于建模用户生命周期阶段。STAN 首先根据学习用户的偏好识别潜在的用户生命周期阶段，然后利用阶段表示来提高多任务学习性能。我们使用公共数据集和工业数据集的实验结果表明，与最先进的方法相比，该模型显著提高了多任务预测性能，突出了在推荐系统中考虑用户生命周期阶段的重要性。在线 A/B 测试表明，我们的模型优于现有的模型，在每个用户的停留时间和 CVR 分别达到了3.05% 和0.88% 的显著改善。我们已经在所有 Shopee 直播推荐服务上部署了 STAN。	code	0
Bootstrapped Personalized Popularity for Cold Start Recommender Systems	Iason Chaimalas, Duncan Martin Walker, Edoardo Gruppi, Benjamin Richard Clark, Laura Toni	UCL, London, England; British Broadcasting Corp, London, England	Recommender Systems are severely hampered by the well-known Cold Start problem, identified by the lack of information on new items and users. This has led to research efforts focused on data imputation and augmentation models as predominantly data preprocessing strategies, yet their improvement of cold-user performance is largely indirect and often comes at the price of a reduction in accuracy for warmer users. To address these limitations, we propose Bootstrapped Personalized Popularity (B2P), a novel framework that improves performance for cold users (directly) and cold items (implicitly) via popularity models personalized with item metadata. B2P is scalable to very large datasets and directly addresses the Cold Start problem, so it can complement existing Cold Start strategies. Experiments on a real-world dataset from the BBC iPlayer and a public dataset demonstrate that B2P (1) significantly improves cold-user performance, (2) boosts warm-user performance for bootstrapped models by lowering their training sparsity, and (3) improves total recommendation accuracy at a competitive diversity level relative to existing high-performing Collaborative Filtering models. We demonstrate that B2P is a powerful and scalable framework for strongly cold datasets.	推荐系统受到众所周知的冷启动问题的严重阻碍，因为缺乏关于新项目和用户的信息。这导致研究工作将重点放在数据估算和增强模型上，将其作为主要的数据预处理战略，但它们对冷用户性能的改善在很大程度上是间接的，而且往往是以降低较温暖用户的准确性为代价的。为了解决这些局限性，我们提出了一种新的引导式个性化流行(Bootstrap Personalization Popular，B2P)框架，该框架通过使用项目元数据个性化的流行模型来提高冷用户(直接)和冷项目(隐式)的性能。B2P 可以扩展到非常大的数据集，并且可以直接解决冷启动问题，因此它可以补充现有的冷启动策略。在来自 BBC iPlayer 的现实数据集和一个公共数据集上的实验表明，B2P (1)显著提高了冷用户的性能，(2)通过降低训练稀疏性提高了自举模型的热用户性能，(3)相对于现有的高性能协同过滤模型，在竞争多样性水平上提高了总体推荐的准确性。我们证明了 B2P 对于强冷数据集是一个强大的、可扩展的框架。	code	0
Co-occurrence Embedding Enhancement for Long-tail Problem in Multi-Interest Recommendation	Yaokun Liu, Xiaowang Zhang, Minghui Zou, Zhiyong Feng	Tianjin Univ, Tianjin, Peoples R China	Multi-interest recommendation methods extract multiple interest vectors to represent the user comprehensively. Despite their success in the matching stage, previous works overlook the long-tail problem. This results in the model excelling at suggesting head items, while the performance for tail items, which make up more than 70% of all items, remains suboptimal. Hence, enhancing the tail item recommendation capability holds great potential for improving the performance of the multi-interest model. Through experimental analysis, we reveal that the insufficient context for embedding learning is the reason behind the under-performance of tail items. Meanwhile, we face two challenges in addressing this issue: the absence of supplementary item features and the need to maintain head item performance. To tackle these challenges, we propose a CoLT module (Co-occurrence embedding enhancement for Long-Tail problem) that replaces the embedding layer of existing multi-interest frameworks. By linking co-occurring items to establish "assistance relationships", CoLT aggregates information from relevant head items into tail item embeddings and enables joint gradient updates. Experiments on three datasets show our method outperforms SOTA models by 21.86% Recall@50 and improves the Recall@50 of tail items by 14.62% on average.	多兴趣推荐方法提取多个兴趣向量来全面表示用户。尽管他们的成功在匹配阶段，以往的作品忽视了长尾问题。这导致模型在建议首项方面表现出色，而尾项(占所有项目的70% 以上)的表现仍然不理想。因此，提高尾部项目推荐能力对于提高多兴趣模型的性能具有很大的潜力。通过实验分析，我们发现嵌入式学习环境的不足是尾项表现不佳的原因。与此同时，我们在解决这一问题时面临两个挑战: 缺乏补充项目特征和需要保持首项目的性能。为了应对这些挑战，我们提出了一个 CoLT 模块(针对 Long-Tail 问题的共现嵌入增强) ，以取代现有多重兴趣框架的嵌入层。CoLT 通过链接共现项目来建立“协助关系”，将相关头项目的信息聚合成尾项目嵌入，并实现联合梯度更新。在三个数据集上的实验表明，该方法比 SOTA 模型提高了21.86% 的召回率@50，平均提高了14.62% 的尾项召回率@50。	code	0
On the Consistency of Average Embeddings for Item Recommendation	Walid Bendada, Guillaume SalhaGalvan, Romain Hennequin, Thomas Bouabça, Tristan Cazenave	Deezer, Paris, France; Univ Paris 09, Deezer, Paris, Dauphine, France; Univ Paris 09, LAMSADE, PSL, Paris, Dauphine, France	A prevalent practice in recommender systems consists of averaging item embeddings to represent users or higher-level concepts in the same embedding space. This paper investigates the relevance of such a practice. For this purpose, we propose an expected precision score, designed to measure the consistency of an average embedding relative to the items used for its construction. We subsequently analyze the mathematical expression of this score in a theoretical setting with specific assumptions, as well as its empirical behavior on real-world data from music streaming services. Our results emphasize that real-world averages are less consistent for recommendation, which paves the way for future research to better align real-world embeddings with assumptions from our theoretical setting.	在推荐系统中，一个流行的实践是在相同的嵌入空间中平均条目嵌入来表示用户或者更高层次的概念。本文探讨了这种实践的相关性。为此，我们提出了一个期望精度得分，设计用来衡量平均嵌入的一致性相对于其构造所使用的项目。随后，我们分析了这个乐谱的数学表达式在一个理论设置与具体的假设，以及它的经验行为对现实世界的数据从音乐流媒体服务。我们的研究结果强调，现实世界的平均值与推荐的一致性较差，这为以后的研究更好地将现实世界的嵌入与我们理论设置的假设结合起来铺平了道路。	code	0
Progressive Horizon Learning: Adaptive Long Term Optimization for Personalized Recommendation	Congrui Yi, David Zumwalt, Zijian Ni, Shreya Chakrabarti	Amazon, Seattle, WA 98109 USA	As E-commerce and subscription services scale, personalized recommender systems are often needed to further drive long term business growth in acquisition, engagement, and retention of customers. However, long-term metrics associated with these goals can require several months to mature. Additionally, deep personalization also demands a large volume of training data that take a long time to collect. These factors incur substantial lead time for training a model to optimize a long-term metric. Before such model is deployed, a recommender system has to rely on a simple policy (e.g. random) to collect customer feedback data for training, inflicting high opportunity cost and delaying optimization of the target metric. Besides, as customer preferences can shift over time, a large temporal gap between inputs and outcome poses a high risk of data staleness and suboptimal learning. Existing approaches involve various compromises. For instance, contextual bandits often optimize short-term surrogate metrics with simple model structure, which can be suboptimal in the long run, while Reinforcement Learning approaches rely on an abundance of historical data for offline training, which essentially means long lead time before deployment. To address these problems, we propose Progressive Horizon Learning Recommender (PHLRec), a personalized model that can progressively learn metric patterns and adaptively evolve from short- to long-term optimization over time. Through simulations and real data experiments, we demonstrated that PHLRec outperforms competing methods, achieving optimality in both deployment speed and long-term metric performances.	随着电子商务和订阅服务规模的扩大，个性化推荐系统往往需要进一步推动长期业务增长的收购，参与和保留客户。然而，与这些目标相关的长期指标可能需要几个月才能成熟。此外，深度个性化还需要大量的培训数据，这些数据需要很长时间才能收集到。这些因素导致培训模型以优化长期度量的大量提前时间。在采用这种模式之前，推荐系统必须依靠一个简单的策略(例如随机)来收集客户反馈数据，以便进行培训，造成较高的机会成本，并延迟目标指标的优化。此外，由于客户偏好可以随时间变化，输入和输出之间的时间差距很大，造成数据过时和次优学习的高风险。现有的方法涉及各种折衷方案。例如，上下文强盗经常使用简单的模型结构优化短期替代指标，从长远来看这可能是次优的，而强化学习方法依赖于大量的离线培训历史数据，这基本上意味着部署前的长时间准备时间。为了解决这些问题，我们提出了渐进式视野学习推荐器(PHLRec) ，这是一个个性化的模型，它可以逐步学习度量模式，并随着时间的推移自适应地从短期优化演变为长期优化。通过仿真和实际数据实验，我们证明了 PHLRec 优于竞争方法，在部署速度和长期指标性能方面都达到了最优。	code	0
From Research to Production: Towards Scalable and Sustainable Neural Recommendation Models on Commodity CPU Hardware	Anshumali Shrivastava, Vihan Lakshman, Tharun Medini, Nicholas Meisburger, Joshua Engels, David Torres Ramos, Benito Geordie, Pratik Pranav, Shubh Gupta, Yashwanth Adunukota, Siddharth Jain	ThirdAI Corp, Houston, TX 77027 USA	In the last decade, large-scale deep learning has fundamentally transformed industrial recommendation systems. However, this revolutionary technology remains prohibitively expensive due to the need for costly and scarce specialized hardware, such as Graphics Processing Units (GPUs), to train and serve models. In this talk, we share our multi-year journey at ThirdAI in developing efficient neural recommendation models that can be trained and deployed on commodity CPU machines without the need for costly accelerators like GPUs. In particular, we discuss the limitations of the current GPU-based ecosystem in machine learning, why recommendation systems are amenable to the strengths of CPU devices, and present results from our efforts to translate years of academic research into a deployable system that fundamentally shifts the economics of training and operating large-scale machine learning models.	在过去的十年里，大规模的深度学习从根本上改变了行业推荐系统。然而，这种革命性的技术仍然昂贵，由于需要昂贵和稀缺的专业硬件，如图形处理单元(GPU) ，培训和服务模型。在这个演讲中，我们分享了我们在 ThirdAI 多年的发展有效的神经推荐模型的旅程，这些模型可以在普通的 CPU 机器上训练和部署，而不需要像 GPU 这样昂贵的加速器。特别是，我们讨论了当前基于 GPU 的机器学习生态系统的局限性，为什么推荐系统适合 CPU 设备的优势，并介绍了我们将多年的学术研究转化为可部署系统的努力的结果，从根本上改变了培训和操作大规模机器学习模型的经济学。	code	0
User-Centric Conversational Recommendation: Adapting the Need of User with Large Language Models	Gangyi Zhang	Univ Sci & Technol China, Hefei, Peoples R China	Conversational recommender systems (CRS) promise to provide a more natural user experience for exploring and discovering items of interest through ongoing conversation. However, effectively modeling and adapting to users' complex and changing preferences remains challenging. This research develops user-centric methods that focus on understanding and adapting to users throughout conversations to provide the most helpful recommendations. First, a graph-based Conversational Path Reasoning (CPR) framework is proposed that represents dialogs as interactive reasoning over a knowledge graph to capture nuanced user interests and explain recommendations. To further enhance relationship modeling, graph neural networks are incorporated for improved representation learning. Next, to address uncertainty in user needs, the Vague Preference Multi-round Conversational Recommendation (VPMCR) scenario and matching Adaptive Vague Preference Policy Learning (AVPPL) solution are presented using reinforcement learning to tailor recommendations to evolving preferences. Finally, opportunities to leverage large language models are discussed to further advance user experiences via advanced user modeling, policy learning, and response generation. Overall, this research focuses on designing conversational recommender systems that continuously understand and adapt to users' ambiguous, complex and changing needs during natural conversations.	会话推荐系统(CRS)承诺通过正在进行的会话为探索和发现感兴趣的项目提供更自然的用户体验。然而，有效地建模和适应用户复杂和不断变化的偏好仍然具有挑战性。这项研究开发了以用户为中心的方法，重点是在整个对话过程中理解和适应用户，以提供最有用的建议。首先，提出了一个基于图的会话路径推理(CPR)框架，该框架将对话表示为知识图上的交互式推理，以获取细微差别的用户兴趣并解释推荐。为了进一步加强关系建模，引入了图神经网络来改进表示学习。接下来，为了解决用户需求中的不确定性，我们提出了 Vague 偏好多轮对话推荐(vPMCR)场景和匹配的自适应 Vague 偏好政策学习(AVPPL)解决方案，使用强化学习来调整推荐以适应不断变化的偏好。最后，讨论了利用大型语言模型的机会，通过高级用户建模、策略学习和响应生成来进一步提升用户体验。总的来说，本研究的重点是设计会话推荐系统，不断理解和适应用户在自然会话过程中模糊、复杂和不断变化的需求。	code	0
Integrating Offline Reinforcement Learning with Transformers for Sequential Recommendation	Xumei Xi, Yuke Zhao, Quan Liu, Liwen Ouyang, Yang Wu		We consider the problem of sequential recommendation, where the current recommendation is made based on past interactions. This recommendation task requires efficient processing of the sequential data and aims to provide recommendations that maximize the long-term reward. To this end, we train a farsighted recommender by using an offline RL algorithm with the policy network in our model architecture that has been initialized from a pre-trained transformer model. The pre-trained model leverages the superb ability of the transformer to process sequential information. Compared to prior works that rely on online interaction via simulation, we focus on implementing a fully offline RL framework that is able to converge in a fast and stable way. Through extensive experiments on public datasets, we show that our method is robust across various recommendation regimes, including e-commerce and movie suggestions. Compared to state-of-the-art supervised learning algorithms, our algorithm yields recommendations of higher quality, demonstrating the clear advantage of combining RL and transformers.	我们考虑顺序推荐的问题，其中当前的推荐是基于过去的交互作用。这项推荐任务需要有效处理顺序数据，目的是提供建议，最大限度地实现长期回报。为此，我们在模型结构中使用策略网络的离线 RL 算法来训练一个有远见的推荐器，该模型已经从一个预先训练好的变压器模型初始化。预先训练的模型利用变压器处理顺序信息的卓越能力。与以往依赖于通过仿真进行在线交互的工作相比，我们侧重于实现一个完全离线的 RL 框架，该框架能够快速、稳定地收敛。通过在公共数据集上的大量实验，我们发现我们的方法在不同的推荐机制下都是稳健的，包括电子商务和电影推荐。与最先进的监督式学习算法相比，我们的算法产生了更高质量的推荐，展示了结合 RL 和变压器的明显优势。	code	0
Fast and Examination-agnostic Reciprocal Recommendation in Matching Markets	Yoji Tomita, Riku Togashi, Yuriko Hashizume, Naoto Ohsaka	CyberAgent Inc, Tokyo, Japan	In matching markets such as job posting and online dating platforms, the recommender system plays a critical role in the success of the platform. Unlike standard recommender systems that suggest items to users, reciprocal recommender systems (RRSs) that suggest other users must take into account the mutual interests of users. In addition, ensuring that recommendation opportunities do not disproportionately favor popular users is essential for the total number of matches and for fairness among users. Existing recommendation methods in matching markets, however, face computational challenges on real-world scale platforms and depend on specific examination functions in the position-based model (PBM). In this paper, we introduce the reciprocal recommendation method based on the matching with transferable utility (TU matching) model in the context of ranking recommendations in matching markets, and propose a faster and examination-agnostic algorithm. Furthermore, we evaluate our approach on experiments with synthetic data and real-world data from an online dating platform in Japan. Our method performs better than or as well as existing methods in terms of the total number of matches and works well even in relatively large datasets for which one existing method does not work.	在招聘和在线约会平台等匹配市场方面，推荐系统对平台的成功起着关键作用。不像标准的推荐系统，建议项目给用户，互惠推荐系统(RRS) ，建议其他用户必须考虑到用户的共同利益。此外，确保推荐机会不会不成比例地偏袒流行用户，对于匹配的总数和用户之间的公平性至关重要。然而，现有的匹配市场推荐方法在实际规模的平台上面临着计算上的挑战，并且依赖于基于位置模型(PBM)中的特定检验函数。本文在匹配市场推荐排序的背景下，介绍了基于匹配可转移效用(TU 匹配)模型的互惠推荐方法，并提出了一种更快、考试无关的算法。此外，我们评估了我们的实验方法与合成数据和真实世界的数据从一个在线约会平台在日本。就匹配总数而言，我们的方法比现有方法执行得更好，甚至在一个现有方法不适用的相对较大的数据集中也能很好地工作。	code	0
✨ Going Beyond Local: Global Graph-Enhanced Personalized News Recommendations	Boming Yang, Dairui Liu, Toyotaro Suzumura, Ruihai Dong, Irene Li				code	0
Distribution-based Learnable Filters with Side Information for Sequential Recommendation	Haibo Liu, Zhixiang Deng, Liang Wang, Jinjia Peng, Shi Feng	HeBei Univ, Sch Cyber Secur & Comp, Baoding, Peoples R China; Northeastern Univ, Sch Comp Sci & Engn, Shenyang, Peoples R China	Sequential Recommendation aims to predict the next item by mining out the dynamic preference from user previous interactions. However, most methods represent each item as a single fixed vector, which is incapable of capturing the uncertainty of item-item transitions that result from time-dependent and multifarious interests of users. Besides, they struggle to effectively exploit side information that helps to better express user preferences. Finally, the noise in user's access sequence, which is due to accidental clicks, can interfere with the next item prediction and lead to lower recommendation performance. To deal with these issues, we propose DLFS-Rec, a simple and novel model that combines Distribution-based Learnable Filters with Side information for sequential Recommendation. Specifically, items and their side information are represented by stochastic Gaussian distribution, which is described by mean and covariance embeddings, and then the corresponding embeddings are fused to generate a final representation for each item. To attenuate noise, stacked learnable filter layers are applied to smooth the fused embeddings. Extensive experiments on four public real-world datasets demonstrate the superiority of the proposed model over state-of-the-art baselines, especially on cold start users and items. Codes are available at https://github.com/zxiang30/DLFS-Rec.	序贯推荐的目的是通过挖掘用户之前交互中的动态偏好来预测下一个项目。然而，大多数方法将每个项目表示为一个单一的固定向量，不能捕捉由于时间依赖性和用户兴趣的多样性而产生的项目-项目转换的不确定性。此外，他们努力有效地利用有助于更好地表达用户偏好的副信息。最后，用户访问序列中由于偶然点击而产生的噪声会干扰下一个项目的预测，从而降低推荐性能。为了解决这些问题，我们提出了 DLFS-Rec 模型，这是一个简单而新颖的模型，它将基于分布的可学习过滤器和侧信息结合起来用于顺序推荐。具体来说，项目和它们的侧面信息用随机正态分布表示，这种表示用均值和协方差嵌入来描述，然后对相应的嵌入进行融合，为每个项目生成最终的表示。为了抑制噪声，采用叠加的可学习滤波层来平滑融合嵌入。在四个公共真实世界数据集上的大量实验表明，该模型优于最先进的基线，特别是在冷启动用户和项目上。密码可在 https://github.com/zxiang30/dlfs-rec 索取。	code	0
Reciprocal Sequential Recommendation	Bowen Zheng, Yupeng Hou, Wayne Xin Zhao, Yang Song, Hengshu Zhu	Renmin Univ China, Gaoling Sch Artificial Intelligence, Beijing, Peoples R China; BOSS Zhipin, Beijing, Peoples R China	Reciprocal recommender system (RRS), considering a two-way matching between two parties, has been widely applied in online platforms like online dating and recruitment. Existing RRS models mainly capture static user preferences, which have neglected the evolving user tastes and the dynamic matching relation between the two parties. Although dynamic user modeling has been well-studied in sequential recommender systems, existing solutions are developed in a user-oriented manner. Therefore, it is non-trivial to adapt sequential recommendation algorithms to reciprocal recommendation. In this paper, we formulate RRS as a distinctive sequence matching task, and further propose a new approach ReSeq for RRS, which is short for Reciprocal Sequential recommendation. To capture dual-perspective matching, we propose to learn fine-grained sequence similarities by co-attention mechanism across different time steps. Further, to improve the inference efficiency, we introduce the self-distillation technique to distill knowledge from the fine-grained matching module into the more efficient student module. In the deployment stage, only the efficient student module is used, greatly speeding up the similarity computation. Extensive experiments on five real-world datasets from two scenarios demonstrate the effectiveness and efficiency of the proposed method. Our code is available at https://github.com/RUCAIBox/ReSeq/.	考虑双方双向匹配的互惠推荐系统已被广泛应用于在线约会和招聘等在线平台。现有的 RRS 模型主要捕捉静态用户偏好，忽略了用户偏好的变化和双方之间的动态匹配关系。尽管动态用户建模已经在顺序推荐系统中得到了很好的研究，但是现有的解决方案都是以面向用户的方式开发的。因此，将顺序推荐算法应用到互惠推荐中具有重要意义。本文将 RRS 作为一个独特的序列匹配任务，并进一步提出了一种新的 RRS 方法 ReSeq，即相互序列推荐的简称。为了捕获双视角匹配，我们提出了通过跨不同时间步长的共注意机制来学习细粒度序列相似性。进一步，为了提高推理效率，我们引入了自蒸馏技术，从细粒度匹配模块中提取知识到更高效的学生模块中。在部署阶段，只使用了有效的学生模块，大大加快了相似度计算的速度。通过对来自两个场景的五个真实世界数据集的大量实验，证明了该方法的有效性和高效性。我们的代码可以在 https://github.com/rucaibox/reseq/找到。	code	0
STRec: Sparse Transformer for Sequential Recommendations	Chengxi Li, Yejing Wang, Qidong Liu, Xiangyu Zhao, Wanyu Wang, Yiqi Wang, Lixin Zou, Wenqi Fan, Qing Li	Hong Kong Polytech Univ, Hong Kong, Peoples R China; Michigan State Univ, E Lansing, MI 48824 USA; Wuhan Univ, Wuhan, Peoples R China; City Univ Hong Kong, Hong Kong, Peoples R China	With the rapid evolution of transformer architectures, researchers are exploring their application in sequential recommender systems (SRSs) and presenting promising performance on SRS tasks compared with former SRS models. However, most existing transformer-based SRS frameworks retain the vanilla attention mechanism, which calculates the attention scores between all item-item pairs. With this setting, redundant item interactions can harm the model performance and consume much computation time and memory. In this paper, we identify the sparse attention phenomenon in transformer-based SRS models and propose Sparse Transformer for sequential Recommendation tasks (STRec) to achieve the efficient computation and improved performance. Specifically, we replace self-attention with cross-attention, making the model concentrate on the most relevant item interactions. To determine these necessary interactions, we design a novel sampling strategy to detect relevant items based on temporal information. Extensive experimental results validate the effectiveness of STRec, which achieves the state-of-the-art accuracy while reducing 54% inference time and 70% memory cost. We also provide massive extended experiments to further investigate the property of our framework.	随着变压器结构的快速发展，研究人员正在探索其在顺序推荐系统(SRS)中的应用，并与以往的 SRS 模型相比，在 SRS 任务中表现出了良好的性能。然而，大多数现有的基于转换器的 SRS 框架保留了普通的注意机制，它计算所有项目-项目对之间的注意得分。通过这种设置，冗余的项目交互会损害模型的性能，并消耗大量的计算时间和内存。针对基于变压器的 SRS 模型中存在的稀疏注意现象，提出了针对序贯推荐任务的稀疏变压器算法(STRec) ，以提高计算效率和性能。具体来说，我们用交叉注意代替自我注意，使模型集中于最相关的项目交互。为了确定这些必要的交互作用，我们设计了一种新的基于时间信息的抽样策略来检测相关项目。大量的实验结果验证了 STRec 算法的有效性，该算法在减少54% 的推理时间和70% 的内存开销的同时，达到了最先进的精度。我们还提供了大量的扩展实验，以进一步研究我们的框架的性质。	code	0
Deep Situation-Aware Interaction Network for Click-Through Rate Prediction	Yimin Lv, Shuli Wang, Beihong Jin, Yisong Yu, Yapeng Zhang, Jian Dong, Yongkang Wang, Xingxing Wang, Dong Wang	Meituan, Beijing, Peoples R China; Univ Chinese Acad Sci, Chinese Acad Sci, Inst Software, Beijing, Peoples R China	User behavior sequence modeling plays a significant role in Click-Through Rate (CTR) prediction on e-commerce platforms. Except for the interacted items, user behaviors contain rich interaction information, such as the behavior type, time, location, etc. However, so far, the information related to user behaviors has not yet been fully exploited. In the paper, we propose the concept of a situation and situational features for distinguishing interaction behaviors and then design a CTR model named Deep Situation-Aware Interaction Network (DSAIN). DSAIN first adopts the reparameterization trick to reduce noise in the original user behavior sequences. Then it learns the embeddings of situational features by feature embedding parameterization and tri-directional correlation fusion. Finally, it obtains the embedding of behavior sequence via heterogeneous situation aggregation. We conduct extensive offline experiments on three real-world datasets. Experimental results demonstrate the superiority of the proposed DSAIN model. More importantly, DSAIN has increased the CTR by 2.70%, the CPM by 2.62%, and the GMV by 2.16% in the online A/B test. Now, DSAIN has been deployed on the Meituan food delivery platform and serves the main traffic of the Meituan takeout app. Our source code is available at https://github.com/W-void/DSAIN.	用户行为序列建模在电子商务平台的点进率预测中扮演着重要的角色。除了交互项，用户行为还包含丰富的交互信息，如行为类型、时间、地点等。然而，到目前为止，与用户行为相关的信息还没有被充分利用。提出了区分交互行为的情境和情境特征的概念，并设计了一个名为深度情境感知交互网络(DSAIN)的 CTR 模型。DSAIN 首先采用重新参数化技巧来降低原始用户行为序列中的噪声。然后通过特征嵌入参量化和三向相关融合学习情景特征的嵌入。最后，通过异构情景聚合得到行为序列的嵌入。我们在三个真实世界的数据集上进行了大量的离线实验。实验结果表明了所提出的 DSAIN 模型的优越性。更重要的是，在线 A/B 测试中，DSAIN 使 CTR 提高了2.70% ，CPM 提高了2.62% ，GMV 提高了2.16% 。现在，DSAIN 已经部署在美团外卖平台上，为美团外卖应用程序的主要流量提供服务。我们的源代码可以在 https://github.com/w-void/dsain 找到。	code	0
Equivariant Contrastive Learning for Sequential Recommendation	Peilin Zhou, Jingqi Gao, Yueqi Xie, Qichen Ye, Yining Hua, Jaeboum Kim, Shoujin Wang, Sunghun Kim				code	0
Task Aware Feature Extraction Framework for Sequential Dependence Multi-Task Learning	Xuewen Tao, Mingming Ha, Qiongxu Ma, Hongwei Cheng, Wenfang Lin, Xiaobo Guo, Linxun Cheng, Bing Han				code	0
AutoOpt: Automatic Hyperparameter Scheduling and Optimization for Deep Click-through Rate Prediction	Yujun Li, Xing Tang, Bo Chen, Yimin Huang, Ruiming Tang, Zhenguo Li				code	0
Alleviating the Long-Tail Problem in Conversational Recommender Systems	Zhipeng Zhao, Kun Zhou, Xiaolei Wang, Wayne Xin Zhao, Fan Pan, Zhao Cao, JiRong Wen				code	0
Reproducibility of Multi-Objective Reinforcement Learning Recommendation: Interplay between Effectiveness and Beyond-Accuracy Perspectives	Vincenzo Paparella, Vito Walter Anelli, Ludovico Boratto, Tommaso Di Noia				code	0
Personalised Recommendations for the BBC iPlayer: Initial approach and current challenges	Benjamin Richard Clark, Kristine Grivcova, Polina Proutskova, Duncan Martin Walker				code	0
Heterogeneous Knowledge Fusion: A Novel Approach for Personalized Recommendation via LLM	Bin Yin, Junjie Xie, Yu Qin, Zixiang Ding, Zhichao Feng, Xiang Li, Wei Lin				code	0
MCM: A Multi-task Pre-trained Customer Model for Personalization	Rui Luo, Tianxin Wang, Jingyuan Deng, Peng Wan				code	0
Beyond the Sequence: Statistics-Driven Pre-training for Stabilizing Sequential Recommendation Model	Sirui Wang, Peiguang Li, Yunsen Xian, Hongzhi Zhang				code	0
Personalized Category Frequency prediction for Buy It Again recommendations	Amit Pande, Kunal Ghosh, Rankyung Park				code	0
Hessian-aware Quantized Node Embeddings for Recommendation	Huiyuan Chen, Kaixiong Zhou, KweiHerng Lai, ChinChia Michael Yeh, Yan Zheng, Xia Hu, Hao Yang				code	0
Scalable Approximate NonSymmetric Autoencoder for Collaborative Filtering	Martin Spisák, Radek Bartyzal, Antonín Hoskovec, Ladislav Peska, Miroslav Tuma				code	0
M3REC: A Meta-based Multi-scenario Multi-task Recommendation Framework	Zerong Lan, Yingyi Zhang, Xianneng Li				code	0
Incorporating Time in Sequential Recommendation Models	Mostafa Rahmani, James Caverlee, Fei Wang				code	0
Enhancing Transformers without Self-supervised Learning: A Loss Landscape Perspective in Sequential Recommendation	Vivian Lai, Huiyuan Chen, ChinChia Michael Yeh, Minghua Xu, Yiwei Cai, Hao Yang				code	0
Initiative transfer in conversational recommender systems	Yuan Ma, Jürgen Ziegler				code	0
RecQR: Using Recommendation Systems for Query Reformulation to correct unseen errors in spoken dialog systems	Manik Bhandari, Mingxian Wang, Oleg Poliannikov, Kanna Shimizu				code	0
Optimizing Podcast Discovery: Unveiling Amazon Music's Retrieval and Ranking Framework	Geetha Sai Aluri, Paul Greyson, Joaquin Delgado				code	0
OutRank: Speeding up AutoML-based Model Search for Large Sparse Data sets with Cardinality-aware Feature Ranking	Blaz Skrlj, Blaz Mramor				code	0
Improving Group Recommendations using Personality, Dynamic Clustering and Multi-Agent MicroServices	Patrícia Alves, André Martins, Paulo Novais, Goreti Marreiros				code	0
Power Loss Function in Neural Networks for Predicting Click-Through Rate	Ergun Biçici				code	0
Sequential Recommendation Models: A Graph-based Perspective	Andreas Peintner				code	0
Retrieval-augmented Recommender System: Enhancing Recommender Systems with Large Language Models	Dario Di Palma				code	0
Leveraging Large Language Models for Sequential Recommendation	Jesse Harte, Wouter Zorgdrager, Panos Louridas, Asterios Katsifodimos, Dietmar Jannach, Marios Fragkoulis				code	0
Uncovering User Interest from Biased and Noised Watch Time in Video Recommendation	Haiyuan Zhao, Lei Zhang, Jun Xu, Guohao Cai, Zhenhua Dong, JiRong Wen				code	0
Nonlinear Bandits Exploration for Recommendations	Yi Su, Minmin Chen				code	0
SPARE: Shortest Path Global Item Relations for Efficient Session-based Recommendation	Andreas Peintner, Amir Reza Mohammadi, Eva Zangerle				code	0
When Fairness meets Bias: a Debiased Framework for Fairness aware Top-N Recommendation	Jiakai Tang, Shiqi Shen, Zhipeng Wang, Zhi Gong, Jingsen Zhang, Xu Chen				code	0
A Probabilistic Position Bias Model for Short-Video Recommendation Feeds	Olivier Jeunen				code	0
Collaborative filtering algorithms are prone to mainstream-taste bias	Pantelis Pipergias Analytis, Philipp Hager				code	0
Providing Previously Unseen Users Fair Recommendations Using Variational Autoencoders	Bjørnar Vassøy, Helge Langseth, Benjamin Kille				code	0
Large Language Models are Competitive Near Cold-start Recommenders for Language- and Item-based Preferences	Scott Sanner, Krisztian Balog, Filip Radlinski, Ben Wedin, Lucas Dixon				code	0
Towards Companion Recommenders Assisting Users' Long-Term Journeys	Konstantina Christakopoulou, Minmin Chen				code	0
How Users Ride the Carousel: Exploring the Design of Multi-List Recommender Interfaces From a User Perspective	Benedikt Loepp, Jürgen Ziegler				code	0
User Behavior Modeling with Deep Learning for Recommendation: Recent Advances	Weiwen Liu, Wei Guo, Yong Liu, Ruiming Tang, Hao Wang				code	0
HUMMUS: A Linked, Healthiness-Aware, User-centered and Argument-Enabling Recipe Data Set for Recommendation	Felix Bölz, Diana Nurbakova, Sylvie Calabretto, Armin Gerl, Lionel Brunie, Harald Kosch				code	0
Data-free Knowledge Distillation for Reusing Recommendation Models	Cheng Wang, Jiacheng Sun, Zhenhua Dong, Jieming Zhu, Zhenguo Li, Ruixuan Li, Rui Zhang				code	0
Contextual Multi-Armed Bandit for Email Layout Recommendation	Yan Chen, Emilian Vankov, Linas Baltrunas, Preston Donovan, Akash Mehta, Benjamin Schroeder, Matthew Herman				code	0
Domain Disentanglement with Interpolative Data Augmentation for Dual-Target Cross-Domain Recommendation	Jiajie Zhu, Yan Wang, Feng Zhu, Zhu Sun				code	0
Reproducibility Analysis of Recommender Systems relying on Visual Features: traps, pitfalls, and countermeasures	Pasquale Lops, Elio Musacchio, Cataldo Musto, Marco Polignano, Antonio Silletti, Giovanni Semeraro				code	0
What We Evaluate When We Evaluate Recommender Systems: Understanding Recommender Systems' Performance using Item Response Theory	Yang Liu, Alan Medlar, Dorota Glowacka				code	0
Identifying Controversial Pairs in Item-to-Item Recommendations	Junyi Shen, Dayvid V. R. Oliveira, Jin Cao, Brian Knott, Goodman Gu, Sindhu Vijaya Raghavan, Yunye Jin, Nikita Sudan, Rob Monarch				code	0
Interpretable User Retention Modeling in Recommendation	Rui Ding, Ruobing Xie, Xiaobo Hao, Xiaochun Yang, Kaikai Ge, Xu Zhang, Jie Zhou, Leyu Lin				code	0
Adversarial Sleeping Bandit Problems with Multiple Plays: Algorithm and Ranking Application	Jianjun Yuan, Wei Lee Woon, Ludovik Coba				code	0
Deliberative Diversity for News Recommendations: Operationalization and Experimental User Study	Lucien Heitz, Juliane A. Lischka, Rana Abdullah, Laura Laugwitz, Hendrik Meyer, Abraham Bernstein				code	0
Group Fairness for Content Creators: the Role of Human and Algorithmic Biases under Popularity-based Recommendations	Stefania Ionescu, Aniko Hannak, Nicolò Pagan				code	0
Scalable Deep Q-Learning for Session-Based Slate Recommendation	Aayush Singha Roy, Edoardo D'Amico, Elias Z. Tragos, Aonghus Lawlor, Neil Hurley				code	0
Optimizing Long-term Value for Auction-Based Recommender Systems via On-Policy Reinforcement Learning	Ruiyang Xu, Jalaj Bhandari, Dmytro Korenkevych, Fan Liu, Yuchen He, Alex Nikulkov, Zheqing Zhu				code	0
Deep Exploration for Recommendation Systems	Zheqing Zhu, Benjamin Van Roy				code	0
Time-Aware Item Weighting for the Next Basket Recommendations	Aleksey Romanov, Oleg Lashinin, Marina Ananyeva, Sergey Kolesnikov				code	0
Multiple Connectivity Views for Session-based Recommendation	Yaming Yang, Jieyu Zhang, Yujing Wang, Zheng Miao, Yunhai Tong				code	0
Navigating the Feedback Loop in Recommender Systems: Insights and Strategies from Industry Practice	Ding Tong, Qifeng Qiao, TingPo Lee, James McInerney, Justin Basilico				code	0
Unleash the Power of Context: Enhancing Large-Scale Recommender Systems with Context-Based Prediction Models	Jan Hartman, Assaf Klein, Davorin Kopic, Natalia Silberstein				code	0
Learning the True Objectives of Multiple Tasks in Sequential Behavior Modeling	Jiawei Zhang				code	0
Analyzing Accuracy versus Diversity in a Health Recommender System for Physical Activities: a Longitudinal User Study	Ine Coppens, Luc Martens, Toon De Pessemier				code	0
EasyStudy: Framework for Easy Deployment of User Studies on Recommender Systems	Patrik Dokoupil, Ladislav Peska				code	0
LLM Based Generation of Item-Description for Recommendation System	Arkadeep Acharya, Brijraj Singh, Naoyuki Onoe				code	0
Exploring Unlearning Methods to Ensure the Privacy, Security, and Usability of Recommender Systems	Jens Leysen				code	0
Complementary Product Recommendation for Long-tail Products	Rastislav Papso				code	0
Challenges for Anonymous Session-Based Recommender Systems in Indoor Environments	Alessio Ferrato				code	0
Recommenders In the wild - Practical Evaluation Methods	Kim Falk, Morten Arngren				code	0
Masked and Swapped Sequence Modeling for Next Novel Basket Recommendation in Grocery Shopping	Ming Li, Mozhdeh Ariannezhad, Andrew Yates, Maarten de Rijke				code	0
Loss Harmonizing for Multi-Scenario CTR Prediction	Congcong Liu, Liang Shi, Pei Wang, Fei Teng, Xue Jiang, Changping Peng, Zhangang Lin, Jingping Shao				code	0
Towards Robust Fairness-aware Recommendation	Hao Yang, Zhining Liu, Zeyu Zhang, Chenyi Zhuang, Xu Chen				code	0
Two-sided Calibration for Quality-aware Responsible Recommendation	Chenyang Wang, Yankai Liu, Yuanqing Yu, Weizhi Ma, Min Zhang, Yiqun Liu, Haitao Zeng, Junlan Feng, Chao Deng				code	0
RecAD: Towards A Unified Library for Recommender Attack and Defense	Changsheng Wang, Jianbai Ye, Wenjie Wang, Chongming Gao, Fuli Feng, Xiangnan He				code	0
Adversarial Collaborative Filtering for Free	Huiyuan Chen, Xiaoting Li, Vivian Lai, ChinChia Michael Yeh, Yujie Fan, Yan Zheng, Mahashweta Das, Hao Yang				code	0
Trending Now: Modeling Trend Recommendations	Hao Ding, Branislav Kveton, Yifei Ma, Youngsuk Park, Venkataramana Kini, Yupeng Gu, Ravi Divvela, Fei Wang, Anoop Deoras, Hao Wang				code	0
A Lightweight Method for Modeling Confidence in Recommendations with Learned Beta Distributions	Norman Knyazev, Harrie Oosterhuis				code	0
Investigating the effects of incremental training on neural ranking models	Benedikt Schifferer, Wenzhe Shi, Gabriel de Souza Pereira Moreira, Even Oldridge, Chris Deotte, Gilberto Titericz, Kazuki Onodera, Praveen Dhinwa, Vishal Agrawal, Chris Green				code	0
Multi-Relational Contrastive Learning for Recommendation	Wei Wei, Lianghao Xia, Chao Huang				code	0
Challenging the Myth of Graph Collaborative Filtering: a Reasoned and Reproducibility-driven Analysis	Vito Walter Anelli, Daniele Malitesta, Claudio Pomo, Alejandro Bellogín, Eugenio Di Sciascio, Tommaso Di Noia				code	0
InTune: Reinforcement Learning-based Data Pipeline Optimization for Deep Recommendation Models	Kabir Nagrecha, Lingyi Liu, Pablo Delgado, Prasanna Padmanabhan				code	0
Generative Learning Plan Recommendation for Employees: A Performance-aware Reinforcement Learning Approach	Zhi Zheng, Ying Sun, Xin Song, Hengshu Zhu, Hui Xiong				code	0
Knowledge-based Multiple Adaptive Spaces Fusion for Recommendation	Meng Yuan, Fuzhen Zhuang, Zhao Zhang, Deqing Wang, Jin Dong				code	0
KGTORe: Tailored Recommendations through Knowledge-aware GNN Models	Alberto Carlo Maria Mancino, Antonio Ferrara, Salvatore Bufi, Daniele Malitesta, Tommaso Di Noia, Eugenio Di Sciascio				code	0
Everyone's a Winner! On Hyperparameter Tuning of Recommendation Models	Faisal Shehzad, Dietmar Jannach				code	0
ADRNet: A Generalized Collaborative Filtering Framework Combining Clinical and Non-Clinical Data for Adverse Drug Reaction Prediction	Haoxuan Li, Taojun Hu, Zetong Xiong, Chunyuan Zheng, Fuli Feng, Xiangnan He, XiaoHua Zhou				code	0
Using Learnable Physics for Real-Time Exercise Form Recommendations	Abhishek Jaiswal, Gautam Chauhan, Nisheeth Srivastava				code	0
ReCon: Reducing Congestion in Job Recommendation using Optimal Transport	Yoosof Mashayekhi, Bo Kang, Jefrey Lijffijt, Tijl De Bie				code	0
Analysis Operations for Constraint-based Recommender Systems	Sebastian Lubos, VietMan Le, Alexander Felfernig, Thi Ngoc Trang Tran				code	0
Generative Next-Basket Recommendation	Wenqi Sun, Ruobing Xie, Junjie Zhang, Wayne Xin Zhao, Leyu Lin, JiRong Wen				code	0
Extended Conversion: Capturing Successful Interactions in Voice Shopping	Elad Haramaty, Zohar S. Karnin, Arnon Lazerson, Liane LewinEytan, Yoelle Maarek				code	0
Widespread Flaws in Offline Evaluation of Recommender Systems	Balázs Hidasi, Ádám Tibor Czapp				code	0
Towards Sustainability-aware Recommender Systems: Analyzing the Trade-off Between Algorithms Performance and Carbon Footprint	Giuseppe Spillo, Allegra De Filippo, Cataldo Musto, Michela Milano, Giovanni Semeraro				code	0
CR-SoRec: BERT driven Consistency Regularization for Social Recommendation	Tushar Prakash, Raksha Jalan, Brijraj Singh, Naoyuki Onoe				code	0
Interface Design to Mitigate Inflation in Recommender Systems	Rana Shahout, Yehonatan Peisakhovsky, Sasha Stoikov, Nikhil Garg				code	0
Towards Self-Explaining Sequence-Aware Recommendation	Alejandro ArizaCasabona, Maria Salamó, Ludovico Boratto, Gianni Fenu				code	0
Ti-DC-GNN: Incorporating Time-Interval Dual Graphs for Recommender Systems	Nikita Severin, Andrey V. Savchenko, Dmitrii Kiselev, Maria Ivanova, Ivan Kireev, Ilya Makarov				code	0
Of Spiky SVDs and Music Recommendation	Darius Afchar, Romain Hennequin, Vincent Guigue				code	0
Topic-Level Bayesian Surprise and Serendipity for Recommender Systems	Tonmoy Hasan, Razvan Bunescu				code	0
Stability of Explainable Recommendation	Sairamvinay Vijayaraghavan, Prasant Mohapatra				code	0
Station and Track Attribute-Aware Music Personalization	M. Jeffrey Mei, Oliver Bembom, Andreas F. Ehmann				code	0
Delivery Hero Recommendation Dataset: A Novel Dataset for Benchmarking Recommendation Algorithms	Yernat Assylbekov, Raghav Bali, Luke Bovard, Christian Klaue				code	0
Creating the next generation of news experience on ekstrabladet.dk with recommender systems	Johannes Kruse, Kasper Lindskow, Michael Riis Andersen, Jes Frellsen				code	0
Leveling Up the Peloton Homescreen: A System and Algorithm for Dynamic Row Ranking	Natalia Chen, Oinam Nganba Meetei, Nilothpal Talukder, Alexey Zankevich				code	0
Uncovering ChatGPT's Capabilities in Recommender Systems	Sunhao Dai, Ninglu Shao, Haiyuan Zhao, Weijie Yu, Zihua Si, Chen Xu, Zhongxiang Sun, Xiao Zhang, Jun Xu				code	0
Continual Collaborative Filtering Through Gradient Alignment	Jaime Hieu Do, Hady W. Lauw				code	0
Broadening the Scope: Evaluating the Potential of Recommender Systems beyond prioritizing Accuracy	Vincenzo Paparella, Dario Di Palma, Vito Walter Anelli, Tommaso Di Noia				code	0
Climbing crags repetitive choices and recommendations	Iustina Ivanova				code	0
Towards Health-Aware Fairness in Food Recipe Recommendation	Mehrdad Rostami, Mohammad Aliannejadi, Mourad Oussalah				code	0
Localify.org: Locally-focus Music Artist and Event Recommendation	Douglas Turnbull, April Trainor, Douglas R. Turnbull, Elizabeth Richards, Kieran Bentley, Victoria Conrad, Paul Gagliano, Cassandra Raineault, Thorsten Joachims				code	0
Re2Dan: Retrieval of Medical Documents for e-Health in Danish	Antonela Tommasel, Rafael PablosSarabia, Ira Assent				code	0
Introducing LensKit-Auto, an Experimental Automated Recommender System (AutoRecSys) Toolkit	Tobias Vente, Michael Ekstrand, Joeran Beel				code	0
Tutorial on Large Language Models for Recommendation	Wenyue Hua, Lei Li, Shuyuan Xu, Li Chen, Yongfeng Zhang				code	0
On Challenges of Evaluating Recommender Systems in an Offline Setting	Aixin Sun				code	0
Trustworthy Recommender Systems: Technical, Ethical, Legal, and Regulatory Perspectives	Markus Schedl, Vito Walter Anelli, Elisabeth Lex				code	0
Customer Lifetime Value Prediction: Towards the Paradigm Shift of Recommender System Objectives	Chuhan Wu, Qinglin Jia, Zhenhua Dong, Ruiming Tang				code	0
Knowledge-Aware Recommender Systems based on Multi-Modal Information Sources	Giuseppe Spillo				code	0
Explainable Graph Neural Network Recommenders; Challenges and Opportunities	Amir Reza Mohammadi				code	0
Overcoming Recommendation Limitations with Neuro-Symbolic Integration	Tommaso Carraro				code	0
Improving Recommender Systems Through the Automation of Design Decisions	Lukas Wegmeth				code	0
Acknowledging Dynamic Aspects of Trust in Recommender Systems	Imane Akdim				code	0
Denoising Explicit Social Signals for Robust Recommendation	Youchen Sun				code	0
Advancing Automation of Design Decisions in Recommender System Pipelines	Tobias Vente				code	0
Demystifying Recommender Systems: A Multi-faceted Examination of Explanation Generation, Impact, and Perception	Giacomo Balloccu				code	0
Enhanced Privacy Preservation for Recommender Systems	Ziqing Wu				code	0
Incentivizing Exploration in Linear Contextual Bandits under Information Gap	Huazheng Wang, Haifeng Xu, Chuanhao Li, Zhiyuan Liu, Hongning Wang				code	0
Ex2Vec: Characterizing Users and Items from the Mere Exposure Effect	Bruno Sguerra, VietAnh Tran, Romain Hennequin				code	0
Accelerating Creator Audience Building through Centralized Exploration	Buket Baran, Guilherme Dinis Junior, Antonina Danylenko, Olayinka S. Folorunso, Gösta Forsum, Maksym Lefarov, Lucas Maystre, Yu Zhao				code	0
Track Mix Generation on Music Streaming Services using Transformers	Walid Bendada, Théo Bontempelli, Mathieu Morlon, Benjamin Chapus, Thibault Cador, Thomas Bouabça, Guillaume SalhaGalvan				code	0
Reward innovation for long-term member satisfaction	Gary Tang, Jiangwei Pan, Henry Wang, Justin Basilico				code	0
Disentangling Motives behind Item Consumption and Social Connection for Mutually-enhanced Joint Prediction	Youchen Sun, Zhu Sun, Xiao Sha, Jie Zhang, Yew Soon Ong				code	0
How Should We Measure Filter Bubbles? A Regression Model and Evidence for Online News	Lien Michiels, Jorre T. A. Vannieuwenhuyze, Jens Leysen, Robin Verachtert, Annelien Smets, Bart Goethals				code	0
Private Matrix Factorization with Public Item Features	Mihaela Curmei, Walid Krichene, Li Zhang, Mukund Sundararajan				code	0
Transparently Serving the Public: Enhancing Public Service Media Values through Exploration	Andreas Grün, Xenija Neufeld				code	0
Evaluating The Effects of Calibrated Popularity Bias Mitigation: A Field Study	Anastasiia Klimashevskaia, Mehdi Elahi, Dietmar Jannach, Lars Skjærven, Astrid Tessem, Christoph Trattner				code	0
An Exploration of Sentence-Pair Classification for Algorithmic Recruiting	Mesut Kaya, Toine Bogers				code	0
RecSys Challenge 2023: Deep Funnel Optimization with a Focus on User Privacy	Rahul Agrawal, Sarang Brahme, Sourav Maitra, Saikishore Kalloori, Abhishek Srivastava, Yong Liu, Athirai A. Irissappane				code	0
Beyond Labels: Leveraging Deep Learning and LLMs for Content Metadata	Saurabh Agrawal, John Trenkle, Jaya Kawale				code	0
Efficient Data Representation Learning in Google-scale Systems	Derek Zhiyuan Cheng, Ruoxi Wang, WangCheng Kang, Benjamin Coleman, Yin Zhang, Jianmo Ni, Jonathan Valverde, Lichan Hong, Ed H. Chi				code	0
The Effect of Third Party Implementations on Reproducibility	Balázs Hidasi, Ádám Tibor Czapp				code	0
Correcting for Interference in Experiments: A Case Study at Douyin	Vivek F. Farias, Hao Li, Tianyi Peng, Xinyuyang Ren, Huawei Zhang, Andrew Zheng				code	0
Visual Representation for Capturing Creator Theme in Brand-Creator Marketplace	Sarel Duanis, Keren Gaiger, Ravid Cohen, Shaked Zychlinski, Asnat GreensteinMessica				code	0
Turning Dross Into Gold Loss: is BERT4Rec really better than SASRec?	Anton Klenitskiy, Alexey Vasilev				code	0
Uncertainty-adjusted Inductive Matrix Completion with Graph Neural Networks	Petr Kasalický, Antoine Ledent, Rodrigo Alves				code	0