用扩散模型解决语言模型在复杂推理中的短板
Beyond Autoregression: Discrete Diffusion for Complex Reasoning and Planning
- 通过多粒度扩散建模，按难易程度优先学习中间目标
- 在计数游戏和数独任务中准确率分别达91.5%和100%
- 无需搜索即可超越传统自回归模型，适合复杂规划任务
自回归语言模型尽管能力强大，但在复杂推理与长期规划任务上表现不佳。本文提出离散扩散模型作为新解决方案。基于子目标不平衡的视角，我们证明扩散模型能有效学习自回归方法难以捕捉的困难子目标。提出多粒度扩散建模（MGDM），在学习过程中根据难度优先处理子目标。在计数游戏（Countdown）、数独（Sudoku）和布尔可满足性问题（Boolean Satisfiability Problems）等复杂任务上，MGDM显著优于自回归模型，且无需使用搜索技术。例如，在计数游戏和数独任务中，准确率分别达到91.5%和100%，而自回归模型仅为45.8%和20.7%。本工作凸显了基于扩散的方法在提升AI复杂语言理解与求解能力方面的潜力。所有代码已开源：https://github.com/HKUNLP/diffusion-vs-ar。
原文摘要 · Abstract (English)
Autoregressive language models, despite their impressive capabilities, struggle with complex reasoning and long-term planning tasks. We introduce discrete diffusion models as a novel solution to these challenges. Through the lens of subgoal imbalance, we demonstrate how diffusion models effectively learn difficult subgoals that elude autoregressive approaches. We propose Multi-Granularity Diffusion Modeling (MGDM), which prioritizes subgoals based on difficulty during learning. On complex tasks like Countdown, Sudoku, and Boolean Satisfiability Problems, MGDM significantly outperforms autoregressive models without using search techniques. For instance, MGDM achieves 91.5\% and 100\% accuracy on Countdown and Sudoku, respectively, compared to 45.8\% and 20.7\% for autoregressive models. Our work highlights the potential of diffusion-based approaches in advancing AI capabilities for sophisticated language understanding and problem-solving tasks. All associated codes are available at \href{https://github.com/HKUNLP/diffusion-vs-ar}{https://github.com/HKUNLP/diffusion-vs-ar}.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。