arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.02730cs.IR
收藏

APAO:面向生成式推荐的自适应前缀感知优化

APAO: Bridging the Training-Inference Gap in Generative Recommendation via Adaptive Prefix-Aware Optimization

DCST, Tsinghua University(清华大学计算机科学与技术系);Quancheng Laboratory(泉城实验室);AIR, Tsinghua University(清华大学智能产业研究院)

Yuanqing Yu, Yifan Wang, Weizhi Ma, Zhiqiang Guo, Min Zhang

AI总结:

针对生成式推荐中训练与推理不一致的问题,提出自适应前缀感知优化框架,通过前缀级损失和自适应最差前缀优化策略,提升模型在束搜索下的候选保留能力。

摘要:

生成式推荐最近作为一种有前景的序列推荐范式出现。它将任务表述为自回归生成过程,根据用户交互历史预测下一个项目的标记。现有的生成式推荐模型通常使用交叉熵损失等标记级似然目标进行训练,同时在推理过程中使用束搜索来生成排序候选。然而,这导致了根本性的训练-推理不一致:标准训练假设真实标记始终可用,而束搜索在推理过程中会剪枝低概率分支,导致当正确项目的前缀得分较低时,该项目被过早丢弃。为了解决这个问题,我们提出了自适应前缀感知优化(APAO)框架,该框架引入前缀级优化损失,以更好地将训练目标与推理设置对齐。此外,我们设计了一种自适应最差前缀优化策略,在训练过程中动态关注最脆弱的前缀,从而增强模型在束搜索约束下保留正确候选的能力。我们提供了理论分析来证明我们框架的有效性和效率。大量实验表明,APAO 持续缓解了训练-推理不一致,并提高了生成式推荐骨干网络的性能。源代码可在 https://github.com/yuyq18/APAO 公开获取。

英文摘要:

Generative recommendation has recently emerged as a promising paradigm for sequential recommendation. It formulates the task as an autoregressive generation process, predicting tokens of the next item conditioned on user interaction histories. Existing generative recommendation models are typically trained with token-level likelihood objectives such as cross-entropy loss, while employing beam search during inference to generate ranked candidates. However, this leads to a fundamental training-inference inconsistency: standard training assumes ground-truth tokens are always available, while beam search prunes low-probability branches during inference, causing the correct item to be prematurely discarded when its prefixes receive low scores. To address this issue, we propose the Adaptive Prefix-Aware Optimization (APAO) framework, which introduces prefix-level optimization losses to better align the training objective with the inference setting. Furthermore, we design an adaptive worst-prefix optimization strategy that dynamically focuses on the most vulnerable prefixes during training, thereby enhancing the model's ability to retain correct candidates under beam search constraints. We provide theoretical analyses to demonstrate the effectiveness and efficiency of our framework. Extensive experiments show that APAO consistently alleviates the training-inference inconsistency and improves performance across generative recommendation backbones. The source code is publicly available at https://github.com/yuyq18/APAO.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑