通过强化学习激励基于神经符号语言的视觉语言推理
Incentivizing Neuro-symbolic Language-based Reasoning in VLMs via Reinforcement Learning
AI总结：
本文基于Qwen3-VL-2B-Instruct模型，通过强化学习提升视觉语言模型的神经符号语言推理能力，在数学、科学和常识问题上准确率提升3.33%，推理令牌减少75%。
摘要：
世界上有7407种语言。但那些世界上不存在的语言呢？人类是否过于狭隘，不关心外星人使用的语言？外星人也是人类！在2016年电影《 Arrival》中，艾米·亚当斯饰演的 linguist 莱斯利·邦克斯博士通过学习非顺序句构成的外星语言（Heptapod）获得了超越时间、预见未来的能力。本文旨在探索神经符号语言中视觉语言概念的表示与推理，并研究在'思考系统'中的分析推理能力和效率的提升。使用Qwen3-VL-2B-Instruct作为基础模型，并利用4个Nvidia H200 GPU节点，本文在包含数学、科学和常识问题的视觉语言评估数据集上实现了3.33%的准确率提升，同时将推理令牌数量减少75%。本文还记录了所面临的计算挑战、扩展可能性以及未来工作，以改进视觉语言模型在神经符号语言中的思考能力。训练和推理设置可在此处找到：https://github.com/i-like-bfs-and-dfs/wolfram-reasoning。
英文摘要：
There are 7,407 languages in the world. But, what about the languages that are not there in the world? Are humans so narrow minded that we don't care about the languages aliens communicate in? Aliens are humans too! In the 2016 movie Arrival, Amy Adams plays a linguist, Dr. Louise Banks who, by learning to think in an alien language (Heptapod) formed of non-sequential sentences, gains the ability to transcend time and look into the future. In this work, I aim to explore the representation and reasoning of vision-language concepts in a neuro-symbolic language, and study improvement in analytical reasoning abilities and efficiency of "thinking systems". With Qwen3-VL-2B-Instruct as base model and 4 $\times$ Nvidia H200 GPU nodes, I achieve an accuracy improvement of 3.33\% on a vision-language evaluation dataset consisting of math, science, and general knowledge questions, while reducing the reasoning tokens by 75\% over SymPy. I've documented the compute challenges faced, scaling possibilities, and the future work to improve thinking in a neuro-symbolic language in vision-language models. The training and inference setup can be found here: https://github.com/i-like-bfs-and-dfs/wolfram-reasoning.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
