生成检索（GR）通过自回归解码目标物品的标识符，在推荐系统中 emerged 为一种有前景的范式。尽管其潜力，当前方法通常依赖于下一个令牌预测模式，将下一个互动物品的每个令牌视为唯一目标。这种狭窄的关注1) 限制了它们捕捉用户偏好细微结构的能力，2) 忽视了解码标识符与用户行为序列之间的深层交互。为应对这些挑战，我们提出了RankGR，一种基于列表级直接偏好优化的生成检索方法，用于推荐。RankGR将检索过程分解为两个互补的阶段：初始评估阶段（IAP）和精细评分阶段（RSP）。在IAP中，我们将一种新的列表级直接偏好优化策略融入GR，从而促进对层次用户偏好的更全面理解以及更有效的偏序建模。RSP然后利用轻量级评分模块，通过与输入序列的交互来细化IAP生成的前λ候选者，从而获得更精确的候选评估。两个阶段在统一的GR模型下联合优化，确保一致性和效率。此外，我们还在训练和部署中实施了若干实用改进，最终实现了一个能够每秒处理近一万次请求的实时系统。在研究和工业数据集上的广泛离线性能，以及在淘宝“Guess You Like”部分的在线收益，验证了RankGR的有效性和可扩展性。
英文摘要：
Generative retrieval (GR) has emerged as a promising paradigm in recommendation systems by autoregressively decoding identifiers of target items. Despite its potential, current approaches typically rely on the next-token prediction schema, which treats each token of the next interacted items as the sole target. This narrow focus 1) limits their ability to capture the nuanced structure of user preferences, and 2) overlooks the deep interaction between decoded identifiers and user behavior sequences. In response to these challenges, we propose RankGR, a Rank-enhanced Generative Retrieval method that incorporates listwise direct preference optimization for recommendation. RankGR decomposes the retrieval process into two complementary stages: the Initial Assessment Phase (IAP) and the Refined Scoring Phase (RSP). In IAP, we incorporate a novel listwise direct preference optimization strategy into GR, thus facilitating a more comprehensive understanding of the hierarchical user preferences and more effective partial-order modeling. The RSP then refines the top-λ candidates generated by IAP with interactions towards input sequences using a lightweight scoring module, leading to more precise candidate evaluation. Both phases are jointly optimized under a unified GR model, ensuring consistency and efficiency. Additionally, we implement several practical improvements in training and deployment, ultimately achieving a real-time system capable of handling nearly ten thousand requests per second. Extensive offline performance on both research and industrial datasets, as well as the online gains on the "Guess You Like" section of Taobao, validate the effectiveness and scalability of RankGR.
Listwise Preference Alignment Optimization for Tail Item Recommendation
Australian Artificial Intelligence Institute (AAII) and School of Compute Sicence, Faculty of Engineering and Information Technology, University of Technology Sydney（澳大利亚人工智能研究所（AAII）和计算机科学学院，工程与信息技术学院，新南威尔士大学）
Mitigating Popularity Bias in Recommendation with Global Listwise Learning and Progressive Bi-Weighting
University of Montreal（蒙特利尔大学）；School of Economics and Management, Beihang University（北京航空航天大学经济管理学院）；College of Computer Science and Artificial Intelligence, Fudan University（复旦大学计算机科学技术学院）；School of Management, Fudan University（复旦大学管理学院）；School of Economics and Management, Tsinghua University（清华大学经济管理学院）；Department of Computer Science and Operations Research, University of Montreal（蒙特利尔大学计算机科学与运筹学系）
Listwise Preference Diffusion Optimization for User Behavior Trajectories Prediction
University of New South Wales（新南威尔士大学）；Macquarie University（麦考瑞大学）；University of California San Diego（加州大学圣地亚哥分校）；Adobe Research（Adobe研究院）；CSIRO’s Data61（CSIRO数据六十一实验室）
SQL-Rank: A Listwise Approach to Collaborative Ranking
University of California, Davis（加利福尼亚大学戴维斯分校）；Department of Computer Science, University of California, Davis（加利福尼亚大学戴维斯分校计算机科学系）；Department of Statistics, University of California, Davis（加利福尼亚大学戴维斯分校统计系）
在本文中，我们提出了一个列表式的方法来构建用户特定的排名在推荐系统中的合作方式。我们将列表方法与以前的逐点和成对方法进行对比，这些方法分别基于将每个评级或每个成对比较作为一个独立的实例。通过扩展（Cao et al. 2007）的工作，我们将列表式协作排名转换为置换模型下的极大似然，该模型基于低秩潜在得分矩阵将概率质量应用于置换。我们提出了一种新的算法称为SQL-Rank，它可以容纳领带和缺失数据，并可以运行在线性时间。我们开发了一个理论框架，用于分析基于置换模型的新表示理论的列表排序方法。应用这个框架的协作排名，我们得出渐近统计率的用户和项目的数量一起增长。我们的结论表明，我们的SQL-Rank方法往往优于目前最先进的算法，如加权MF和BPR的隐式反馈，并取得了良好的效果相比，显式反馈算法，如矩阵分解和协作排名。
