# ReSAIL: Mitigating Collapse in Iterative Agent Self-Distillation

arXiv: https://arxiv.org/abs/2609.39306 · HTML: https://mlsanity.com/paper/2609.39306/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agent skill learning and evolution, one of 16 directions in LLM agents and tool use.
- Builds on 20 of its 25 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Placed in

- LLM agents and tool use / [Agent skill learning and evolution](https://mlsanity.com/agents/agent-skill-learning-and-evolution/)
- Knowledge distillation / [On-policy and self-distillation for LLMs](https://mlsanity.com/distill/on-policy-and-self-distillation-for-llms/)

## Builds on (most cited first, 20 of 20)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [ALFWorld: Aligning Text and Embodied Environments for Interactive Learning](https://mlsanity.com/paper/2010.03768/) (2020)
- [ExpeL: LLM Agents Are Experiential Learners](https://mlsanity.com/paper/2308.10144/) (2023)
- [Group-in-Group Policy Optimization for LLM Agent Training](https://mlsanity.com/paper/2505.10978/) (2025)
- [Scaling Relationship on Learning Mathematical Reasoning with Large Language Models](https://mlsanity.com/paper/2308.01825/) (2023)
- [RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning](https://mlsanity.com/paper/2504.20073/) (2025)
- [Reinforcement Learning via Self-Distillation](https://mlsanity.com/paper/2601.20802/) (2026)
- [Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction](https://mlsanity.com/paper/2412.04454/) (2024)
- [Self-Distillation Enables Continual Learning](https://mlsanity.com/paper/2601.19897/) (2026)
- [ADaPT: As-Needed Decomposition and Planning with Language Models](https://mlsanity.com/paper/2311.05772/) (2023)
- [Android in the Zoo: Chain-of-Action-Thought for GUI Agents](https://mlsanity.com/paper/2403.02713/) (2024)
- [Skill-SD: Skill-Conditioned Self-Distillation for Multi-turn LLM Agents](https://mlsanity.com/paper/2604.10674/) (2026)
- [Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training](https://mlsanity.com/paper/2501.11425/) (2025)
- [Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops](https://mlsanity.com/paper/2607.07663/) (2026)
- [Re-ReST: Reflection-Reinforced Self-Training for Language Agents](https://mlsanity.com/paper/2406.01495/) (2024)
- [Online Experiential Learning for Language Models](https://mlsanity.com/paper/2603.16856/) (2026)
- [Rethinking Continual Experience Internalization for Self-Evolving LLM Agents](https://mlsanity.com/paper/2606.04703/) (2026)
- [Sample-Efficient Learning from Agent Experience](https://mlsanity.com/paper/2607.21051/) (2026)

## Abstract (arXiv)

Iterative self-distillation enables LLM agents to learn from successive deployments, offering a path toward recursive self-improvement (RSI). Yet our experiments with existing methods reveal a collapse in deployment performance across cycles, while task performance with privileged information (PI) also declines. We address this collapse by prioritizing informative interaction steps for distillation and preserving PI-conditioned behavior as the student becomes the next teacher. We introduce Retentive and Selective Augmentation for Iterative Self-Distillation (ReSAIL), a plug-in augmentation for iterative PI-based self-distillation. ReSAIL selects interaction steps where PI most strongly changes the teacher's predictions and balances the resulting distillation losses across trajectories. It also regularizes the student's PI-conditioned output distributions toward those of the frozen teacher at selected and unselected steps to preserve PI-conditioned behavior for supervision in the next cycle. On ALFWorld and TextCraft, ReSAIL sustains substantial gains across model scales over three cycles, with an average absolute gain of 22.5% in final-cycle success rates when added to self-distillation baselines. Sensitivity-guided selection of offline data also improves action prediction accuracy for multimodal GUI agents on AITZ. These findings provide the first evidence that a more robust learning mechanism can effectively mitigate performance collapse in iterative agent self-distillation over deployment trajectories.