跳到正文
arXiv · cs.AI · Rongfeng Guo, Zhuang Zhuang, Ji Dai, Peng Zhao, Shijie Li, Jie Chen, Teng Long, Vincent Tao Hu, Fei Pan · 8 小时前模型评分28

SA-OPSD:面向生成式推荐的搜索感知在线策略自蒸馏框架

Search-Aware On-Policy Self-Distillation with Sibling Alignment for Generative Recommendation

导读

研究者提出搜索感知在线策略自蒸馏框架(SA-OPSD,Search-Aware On-Policy Self-Distillation),把自教师知识迁移到学生束搜索实际访问的状态上。

生成六维分析
分类置信度

中 · 待人工复核

  • 只有单一媒体来源,需要人工复核。

来源级别:媒体与个人

事件类型:研究发布

证据摘录

  • “To address these challenges, we propose Search-Aware On-Policy Self-Distillation (SA-OPSD), a framework that transfers privileged self-teacher knowledge to the states visited during the student's own beam search.”

来源:arXiv · cs.AI · arxiv.org