From Proprietary to Open-Source: Bridging the Distribution Gap via Multi-Agent Protocol Distillation in Agentic Search

• URL: http://arxiv.org/abs/2607.24280v1
• Date: Mon, 27 Jul 2026 11:27:38 GMT
• Title: From Proprietary to Open-Source: Bridging the Distribution Gap via Multi-Agent Protocol Distillation in Agentic Search
• Abstract summary: Agentic search enables large language models to solve knowledge-intensive tasks by interleaving multi-step reasoning with retrieval.<n>Knowledge distillation can supply guidance, and advanced proprietary models with their strong reasoning capabilities are promising teachers.<n>We propose Multi-Agent Protocol Distillation (MAPD), a joint distillation and outcome-based reinforcement learning framework.
• Score: 12.160113054571356
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Agentic search enables large language models to solve knowledge-intensive tasks by interleaving multi-step reasoning with retrieval, yet optimizing this with outcome-based reinforcement learning (RL) provides only sparse supervision. Knowledge distillation can supply denser guidance, and advanced proprietary models with their strong reasoning capabilities are promising teachers. While distilling from proprietary models can densify this supervisory signal, conventional logit-matching is precluded by hidden logits and mismatched tokenizers, whereas raw natural language trajectory imitation transfers superficial stylistic artifacts rather than core reasoning competence. To address the heterogeneous distillation problem and bridge the distribution gap, we propose Multi-Agent Protocol Distillation (MAPD), a joint distillation and RL framework uses a structured, style-normalized protocol as an intermediate representation. An offline multi-agent system (MAS) decomposes each query, retrieves supporting evidence, repairs failed searches, and converts the resulting exploration trace into a JSON protocol containing the task type, reasoning plan, and extractive grounding facts. During training, the protocol is provided only to a privileged branch of the student policy, whose token distributions furnish a dense distillation signal alongside the sparse RL objective. Extensive evaluations across seven QA benchmarks demonstrate that MAPD consistently outperforms competitive distillation and RL, achieving average success rates of 39.4\% on Qwen3-1.7B and 44.4\% on Qwen3-4B. Crucially, the framework generalizes robustly across diverse proprietary teachers while effectively mitigating the student policy from style drift and verbosity degeneration.

Related papers

• Graph Is the Verifier: Agentic Reinforcement Learning for Interprocedural Vulnerability Detection [13.409515289348532]
  VulAgentRL is a framework for interprocedural vulnerability detection built on a Code Property Graph (CPG)<n>We show that VulAgentRL outperforms state-of-the-art baselines, including frontier models, on the strict pair-wise-correct metric while issuing fewer tool calls.
  arXiv  Detail & Related papers  (2026-07-29T09:16:54Z)
• DiffusionOPD: A Unified Perspective of On-Policy Distillation in Diffusion Models [55.01951088768769]
  DiffusionOPD is a new multi-task training paradigm for diffusion models based on Online Policy Distillation (OPD)<n>We show that DiffusionOPD consistently surpasses both multi-reward RL and cascade RL baselines in training efficiency and final performance.
  arXiv  Detail & Related papers  (2026-05-14T16:49:09Z)
• AutoResearch-RL: Perpetual Self-Evaluating Reinforcement Learning Agents for Autonomous Neural Architecture Discovery [5.110708177092157]
  We present AutoResearch-RL, a framework in which a reinforcement learning agent conducts open-ended neural architecture research without human supervision.<n>We formalise this as a Markov Decision Process, derive convergence guarantees under mild assumptions, and demonstrate empirically on a single GPU nanochat pretraining benchmark.
  arXiv  Detail & Related papers  (2026-03-07T17:49:44Z)
• Reinforcement-aware Knowledge Distillation for LLM Reasoning [63.53679456364683]
  Reinforcement learning (RL) post-training has recently driven gains in long chain-of-thought reasoning large language models (LLMs)<n>Most existing knowledge distillation methods are designed for supervised fine-tuning (SFT), relying on fixed teacher traces or teacher-student Kullback-Leibler (KL) divergence-based regularization.<n>We propose RL-aware distillation (RLAD), which performs selective imitation during RL -- guiding the student toward the teacher only when it improves the current policy update.
  arXiv  Detail & Related papers  (2026-02-26T00:20:39Z)
• Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration [49.9937230730202]
  We propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention.<n>Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories.<n>We show that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales.
  arXiv  Detail & Related papers  (2026-02-03T15:32:09Z)
• Coupled Variational Reinforcement Learning for Language Model General Reasoning [83.82392089177841]
  We propose textitbCoupled bVari bReinforcement bLearning (CoVRL) to bridge variational inference and reinforcement learning.<n>CoVRL improves performance by 12.4% over the base model and achieves an additional 2.3% improvement over strong state-of-the-art verifier-free RL baselines.
  arXiv  Detail & Related papers  (2025-12-14T07:03:51Z)
• Multimodal Reinforcement Learning with Agentic Verifier for AI Agents [131.46008226323423]
  Argos is a principled multimodal reward agent to train reasoning models for agentic tasks.<n>By leveraging our agentic verifier across both SFT data and RL training, our model achieves state-of-the-art results.
  arXiv  Detail & Related papers  (2025-12-03T04:42:47Z)
• MALT: Improving Reasoning with Multi-Agent LLM Training [67.76186488361685]
  MALT (Multi-Agent LLM Training) is a novel post-training strategy that divides the reasoning process into generation, verification, and refinement steps.<n>On MATH, GSM8K, and CSQA, MALT surpasses the same baseline LLM with a relative improvement of 15.66%, 7.42%, and 9.40% respectively.
  arXiv  Detail & Related papers  (2024-12-02T19:30:36Z)
• Deep Multi-Agent Reinforcement Learning for Decentralized Active Hypothesis Testing [11.639503711252663]
  We tackle the multi-agent active hypothesis testing (AHT) problem by introducing a novel algorithm rooted in the framework of deep multi-agent reinforcement learning. We present a comprehensive set of experimental results that effectively showcase the agents' ability to learn collaborative strategies and enhance performance.
  arXiv  Detail & Related papers  (2023-09-14T01:18:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.