arXiv cs.AI· Hongbin Lin, Sizhe Zou, Juangui Xu, Xinyue Xu, Zhuoli Ouyang, Stanislav Kolchin, Chengwei Qin, Zhongxiang Dai, Yao Shu·· 10 小时前AI 评分32
DR-BOLT:面向自适应 VLM 推理的双参考离线优化
Learning When to Think: Dual-Reference Offline Optimization for Adaptive VLM Reasoning
AI 导读
针对 VLM 强制长推理导致过度思考的问题,研究者提出双参考离线优化方法 DR-BOLT,区分快/慢思考的混合参考分布与采样分布,并在奖励加权训练中校正二者偏差。
正文 · 原文
Abstract:Reasoning traces improve vision-language models on complex multimodal tasks, but forcing every input to follow a long reasoning process leads to overthinking and unnecessary inference cost. We study adaptive fast/slow thinking for VLMs, where one model answers directly for simple inputs and reasons step by step when needed. Fast answers and slow traces often come from distinct behavior policies, making a single-reference offline objective misweight one mode. We propose DUAL-REFERENCE BOLT (DR-BOLT), which distinguishes the desired fast/slow mixture reference from the sampling distribution and corrects their mismatch in reward-weighted training. Fast/slow partition competition explains routing in the optimal target; an efficient source-wise correction approximates the exact objective when response formats are well separated. Matched Qwen3-VL comparisons use identical candidate banks to isolate the objective from slow-teacher quality. DR-BOLT improves score-length trade-offs over the evaluated offline and online baselines on ScienceQA and three held-out transfer benchmarks, with a 24.1% cross-benchmark slow ratio and a 2.90-fold end-to-end training speedup over reproduced GRPO. These results support adaptive reasoning in the studied two-reference setting, while broader robustness remains open. Our code is available at this https URL.
| Comments: | Accepted to NeurIPS 2026 |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.09154 [cs.AI] |
| (or arXiv:2611.09154v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.09154 arXiv-issued DOI via DataCite |
Submission history
From: Sizhe Zou [view email]
[v1]
Fri, 6 Nov 2026 08:35:20 UTC (1,305 KB)
[v2]
Wed, 18 Nov 2026 09:04:18 UTC (1,305 KB)
来源:arXiv cs.AI · arxiv.org