用扩散语言模型提升搜索代理效率，解决推理慢和工具调用弱的问题。
DLLM-Searcher: Adapting Diffusion Large Language Model for Search Agents
- 设计两阶段微调，增强模型的推理与工具调用能力。
- 提出并行推理新范式，使模型等待工具返回时仍可继续思考。
- 在保持性能的同时，推理速度提升约15%，适合高效搜索场景。
近年来，扩散大语言模型（dLLMs）因其固有的并行解码机制和灵活生成范式，展现出独特的高效性优势。然而，尽管搜索代理发展迅速，其实际部署仍受限于两大挑战：一是延迟挑战——在ReAct范式下，多轮推理、工具调用及等待响应的串行执行导致端到端延迟严重；二是代理能力挑战——现有dLLM在推理和工具调用方面表现薄弱，难以发挥其潜力。本文提出DLLM-Searcher，一种面向dLLM的搜索代理优化框架。为解决代理能力挑战，设计两阶段后训练流程：包含代理监督微调（Agentic SFT）和代理方差缩减偏好优化（Agentic VRPO），显著提升模型的信息检索与推理能力。为缓解延迟挑战，利用dLLM灵活生成特性，提出新型代理范式——并行推理与行动（P-ReAct），引导模型优先解码工具调用指令，实现等待工具返回期间持续推理。实验表明，DLLM-Searcher性能接近主流LLM基搜索代理，且P-ReAct带来约15%的推理加速。代码已开源。
原文摘要 · Abstract (English)
Recently, Diffusion Large Language Models (dLLMs) have demonstrated unique efficiency advantages, enabled by their inherently parallel decoding mechanism and flexible generation paradigm. Meanwhile, despite the rapid advancement of Search Agents, their practical deployment is constrained by a fundamental limitation, termed as 1) Latency Challenge: the serial execution of multi-round reasoning, tool calling, and tool response waiting under the ReAct agent paradigm induces severe end-to-end latency. Intuitively, dLLMs can leverage their distinctive strengths to optimize the operational efficiency of agents under the ReAct agent paradigm. Practically, existing dLLM backbones face the 2) Agent Ability Challenge. That is, existing dLLMs exhibit remarkably weak reasoning and tool-calling capabilities, preventing these advantages from being effectively realized in practice. In this paper, we propose DLLM-Searcher, an optimization framework for dLLM-based Search Agents. To solve the Agent Ability Challenge, we design a two-stage post-training pipeline encompassing Agentic Supervised Fine-Tuning (Agentic SFT) and Agentic Variance-Reduced Preference Optimization Agentic VRPO, which enhances the backbone dLLM's information seeking and reasoning capabilities. To mitigate the Latency Challenge, we leverage the flexible generation mechanism of dLLMs and propose a novel agent paradigm termed Parallel-Reasoning and Acting P-ReAct. P-ReAct guides the model to prioritize decoding tool_call instructions, thereby allowing the model to keep thinking while waiting for the tool's return. Experimental results demonstrate that DLLM-Searcher achieves performance comparable to mainstream LLM-based search agents and P-ReAct delivers approximately 15% inference acceleration. Our code is available at https://anonymous.4open.science/r/DLLM-Searcher-553C
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。