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 2603.14635cs.IRcs.AI
收藏

为推理密集型检索代理进行计算分配

Compute Allocation for Reasoning-Intensive Retrieval Agents

Sreeja Apparaju, Nilesh Gupta

AI总结:

研究通过BRIGHT基准和Gemini 2.5模型家族,探讨推理密集型检索管道中的计算分配,发现重排阶段受益于更强模型和更深候选池,而查询扩展在轻量模型后效果递减。

摘要:

随着代理在长时间范围内运行,其记忆存储持续增长,使检索成为获取相关信息的关键。许多代理查询需要推理密集型检索,其中查询与相关文档之间的联系是隐含的,需要推理来建立连接。LLM增强的管道通过查询扩展和候选重排来解决这个问题,但引入了显著的推理成本。我们使用BRIGHT基准和Gemini 2.5模型家族研究推理密集型检索管道中的计算分配。我们改变模型容量、推理时间思考和重排深度,分别在查询扩展和重排阶段进行调整。我们发现重排阶段从更强的模型(+7.5 NDCG@10)和更深的候选池(从k=10到100增加21%)中受益显著,而查询扩展在轻量模型后效果递减(+1.1 NDCG@10从弱到强)。推理时间思考在两个阶段提供最小改进。这些结果表明,计算应集中在重排阶段,而不是在管道各阶段均匀分布。

英文摘要:

As agents operate over long horizons, their memory stores grow continuously, making retrieval critical to accessing relevant information. Many agent queries require reasoning-intensive retrieval, where the connection between query and relevant documents is implicit and requires inference to bridge. LLM-augmented pipelines address this through query expansion and candidate re-ranking, but introduce significant inference costs. We study computation allocation in reasoning-intensive retrieval pipelines using the BRIGHT benchmark and Gemini 2.5 model family. We vary model capacity, inference-time thinking, and re-ranking depth across query expansion and re-ranking stages. We find that re-ranking benefits substantially from stronger models (+7.5 NDCG@10) and deeper candidate pools (+21% from $k$=10 to 100), while query expansion shows diminishing returns beyond lightweight models (+1.1 NDCG@10 from weak to strong). Inference-time thinking provides minimal improvement at either stage. These results suggest that compute should be concentrated on re-ranking rather than distributed uniformly across pipeline stages.

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。

↑