arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.07359cs.AI
收藏

W&D:通过并行工具调用提升深度研究代理的效率

W&D:Scaling Parallel Tool Calling for Efficient Deep Research Agents

Salesforce AI Research(Salesforce人工智能研究)

Xiaoqiang Lin, Jun Hao Liew, Silvio Savarese, Junnan Li

AI总结:

W&D 通过并行工具调用提升深度研究代理效率,实现性能提升和减少回合数。

摘要:

深度研究代理已经 emerged 为通过多步骤推理和基于网络的信息搜索自动化复杂智力任务的强大工具。尽管最近的努力通过增加连续思考和工具调用的数量来增强这些代理,但通过并行工具调用扩展宽度的潜力仍 largely 未被探索。在本工作中,我们提出了 Wide and Deep 研究代理,一个框架旨在研究通过并行工具调用扩展深度和宽度时代理的行为和性能。与现有方法依赖复杂的多代理协调来并行化工作负载不同,我们的方法利用内在的并行工具调用来在单个推理步骤中实现有效的协调。我们证明了扩展宽度在深度研究基准上显著提高了性能,同时减少了获得正确答案所需的回合数。此外,我们通过案例研究分析了推动这些改进的因素,并探索了各种工具调用调度器以优化并行工具调用策略。我们的发现表明,优化宽度和深度之间的权衡是实现高效深度研究代理的关键路径。值得注意的是,在不使用上下文管理或其他技巧的情况下,我们在 BrowseComp 上使用 GPT-5-Medium 获得了 62.2% 的准确率,超过了 GPT-5-High 原来的 54.9% 准确率。

英文摘要:

Deep research agents have emerged as powerful tools for automating complex intellectual tasks through multi-step reasoning and web-based information seeking. While recent efforts have successfully enhanced these agents by scaling depth through increasing the number of sequential thinking and tool calls, the potential of scaling width via parallel tool calling remains largely unexplored. In this work, we propose the Wide and Deep research agent, a framework designed to investigate the behavior and performance of agents when scaling not only depth but also width via parallel tool calling. Unlike existing approaches that rely on complex multi-agent orchestration to parallelize workloads, our method leverages intrinsic parallel tool calling to facilitate effective coordination within a single reasoning step. We demonstrate that scaling width significantly improves performance on deep research benchmarks while reducing the number of turns required to obtain correct answers. Furthermore, we analyze the factors driving these improvements through case studies and explore various tool call schedulers to optimize parallel tool calling strategy. Our findings suggest that optimizing the trade-off between width and depth is a critical pathway toward high-efficiency deep research agents. Notably, without context management or other tricks, we obtain 62.2% accuracy with GPT-5-Medium on BrowseComp, surpassing the original 54.9% reported by GPT-5-High.

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。

↑