Reranking, as the final stage of recommender systems, plays a crucial role in determining the final exposure, directly influencing user experience. Recently, generative reranking has gained increasing attention for formulating reranking as a holistic sequence generation task, implicitly modeling complex dependencies among items. However, most existing methods suffer from the likelihood trap, where high-likelihood sequences are often repetitive and perceived as low-quality by humans, thereby limiting user engagement. In this work, we propose Consistent Graph-structured Generative Recommendation (CONGRATS). We first introduce a novel Graph-structured Model, which enables the generation of more diverse sequences by exploring multiple paths. This design not only expands the decoding space to promote diversity, but also improves prediction accuracy by explicitly modeling item dependencies from graph transitions. Furthermore, we design a Consistent Differentiable Training method that incorporates an evaluator, allowing the model to learn directly from user preferences. Extensive offline experiments validate the superior performance of CONGRATS over state-of-the-art reranking methods. Moreover, CONGRATS has been evaluated on a large-scale video-sharing app, Kuaishou, with over 300 million daily active users, demonstrating that our approach significantly improves both recommendation quality and diversity, validating our effectiveness in practical industrial platforms.

翻译：重排序作为推荐系统的最终环节，在决定最终曝光内容方面起着至关重要的作用，直接影响用户体验。近年来，生成式重排序因将重排序任务构建为整体序列生成问题，从而隐式建模物品间复杂依赖关系而受到越来越多的关注。然而，现有方法大多受困于“似然陷阱”，即高似然序列往往存在重复性，在人类感知中质量较低，从而限制了用户参与度。本文提出一致性图结构生成式推荐方法（CONGRATS）。我们首先引入一种新颖的图结构模型，该模型能够通过探索多条路径生成更多样化的序列。这一设计不仅扩展了解码空间以促进多样性，还通过显式建模图转移中的物品依赖关系提高了预测准确性。此外，我们设计了一种包含评估器的一致性可微分训练方法，使模型能够直接从用户偏好中学习。大量离线实验验证了CONGRATS相较于最先进重排序方法的优越性能。此外，CONGRATS已在拥有超过3亿日活跃用户的大型视频分享应用快手平台上进行评估，结果表明我们的方法能显著提升推荐质量与多样性，验证了其在实用工业平台上的有效性。