| title | Agent Explorative Policy Optimization for Multimodal Agentic Reasoning |
|---|---|
| authors | Minki Kang, Shizhe Diao, Ryo Hachiuma, Sung Ju Hwang, Pavlo Molchanov, Yu-Chiang Frank Wang, Byung-Kwan Lee |
| date | 2026-05-27 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.28774 |
| affiliation | NVIDIA |
| upvotes | 89 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Vision-language models with extended reasoning succeed on complex problems, but many real-world problems require external tools that internal reasoning alone often cannot resolve. Agentic reasoning therefore interleaves two behaviors with a structural asymmetry: thinking (the self-contained default) and tool use (a high-variance auxiliary acting). We refer to this asymmetry as the Thinking-Acting Gap. Under standard RL recipes like GRPO, the gap manifests as two diagnostic symptoms during training: tool use is attempted on only ~30% of rollouts, and when attempted, the tool-using rollouts within a group are all-wrong on ~40% of questions, suppressing the learning signal at the tool calls that needed it. We propose AXPO (Agent eXplorative Policy Optimization): for each all-wrong tool-using subgroup, AXPO fixes the thinking prefix and resamples the tool call and its continuation, paired with uncertainty-based prefix selection. Across nine multimodal benchmarks and three scales of Qwen3-VL-Thinking, SFT+AXPO outperforms SFT+GRPO at average (+1.8pp Pass@1 and +1.8pp Pass@4 at 8B on average) and 8B with SFT+AXPO surpasses the 32B Base on Pass@4 with 4 times fewer parameters. |
cite
@misc{kang2026agent,
title={Agent Explorative Policy Optimization for Multimodal Agentic Reasoning},
author={Minki Kang and Shizhe Diao and Ryo Hachiuma and Sung Ju Hwang and Pavlo Molchanov and Yu-Chiang Frank Wang and Byung-Kwan Lee},
year={2026},
eprint={2605.28774},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.28774}
}