让大模型搜索时更会取舍，避免信息过载。
Adaptive Information Control for Search-Augmented LLM Reasoning
- 用信息价值动态判断是否继续检索、暴露多少细节。
- 在7个任务中比现有方法平均提升8.6至9.4分。
- 训练后无需外部控制，适合真实场景推理应用。
搜索增强型推理代理在多步推理中交替使用外部检索，但不受控的检索可能引入冗余证据、塞满上下文并破坏强化学习（RL）稳定性。现有基于结果的RL方法仅提供稀疏的终局奖励，难以指导中间信息获取决策。我们提出DeepControl，一种基于信息效用的自适应信息控制框架，该效用为状态相关的检索证据边际价值估计。该框架沿两个维度调控信息获取：广度（是否继续检索）和粒度（暴露多少细节）。通过检索续接引导、层级粒度控制及退火式控制强制机制实现。使策略在训练中内化高效获取行为，测试时无需外部控制。在七个基准上，DeepControl持续优于强基线，未显式控制的情况下，相比Search-R1，在Qwen2.5-7B和Qwen2.5-3B上分别提升+9.4和+8.6分。额外分析显示搜索效率、训练稳定性和证据利用率均提升。
原文摘要 · Abstract (English)
Search-augmented reasoning agents interleave multi-step reasoning with external retrieval, but uncontrolled retrieval can introduce redundant evidence, saturate the context, and destabilize reinforcement learning (RL). Existing outcome-based RL methods provide only sparse terminal rewards, offering limited guidance for intermediate information-acquisition decisions. We propose DeepControl, an adaptive information-control framework based on information utility, a state-dependent estimate of the marginal value of retrieved evidence. The framework regulates information acquisition along two axes: extent, i.e., whether retrieval should continue, and resolution, i.e., how much retrieved detail should be exposed. It implements these controls through retrieval-continuation guidance, hierarchical granularity control, and an annealed control-forcing scheme. This enables the policy to internalize effective acquisition behavior during training and operate without external control at test time. Across seven benchmarks, DeepControl consistently outperforms strong RL and retrieval baselines without explicit information control; compared with Search-R1, it improves average performance by +9.4 and +8.6 points on Qwen2.5-7B and Qwen2.5-3B, respectively. Additional analyses show improved search effectiveness, training stability, and evidence utilization.
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。