Diffusion language models improve plan generation and repair efficiency
Plan-and-Patch: Diffusion Language Models for Agentic Planning
Artificial IntelligenceComputation and LanguageMachine Learning
Summary
Planning long tasks is hard because things can go wrong or change along the way. The authors present a method where a special kind of AI model, called a diffusion language model, helps create and fix plans by only changing the parts that need fixing instead of starting over. This makes planning faster and more reliable. They show their approach works better than traditional methods in tests without special training and is faster after training on task-specific data.
What this means in practice
- •For robotics developers: Improve robots’ ability to quickly create and fix multi-step action plans when tasks or environments change unexpectedly.
- •For game ai programmers: Enable non-player characters to generate and adapt complex game strategies faster without fully redoing plans after failures.
Authors
Syamantak Kumar, Jiang Guo, Hassan Hamad, Hideo Kobayashi, Yi Xiang, Yezhou Yang, Yanjun Qi, Daniele Bonadiman, Jiarong Jiang
Abstract
Planning is increasingly important for long-horizon agents, where successful execution requires coordinating subgoals, tool use, and intermediate outcomes over many steps. Yet assumptions made during planning may be invalidated by the environment, tools may return unexpected results, or actions may fail. Effective agents must therefore not only generate plans, but also revise them. Such revisions often affect only part of a plan, leaving the preceding and subsequent structure intact. Rather than regenerate the entire plan and risk unnecessary changes, repair can regenerate the affected region conditioned on the preserved prefix and suffix. We introduce Plan-and-Patch, a plan-and-act framework in which a diffusion language model (dLLM) generates a structured, program-like plan through parallel unmasking and repairs it by filling in selected regions while keeping the surrounding steps fixed. We compare DreamReasoner-8B and Qwen3-8B as diffusion and autoregressive (AR) planners. On Natural Plan without task-specific training, diffusion (53.7%) achieves nearly twice the plan repair success rate of AR (27.0%). After task-specific training on agentic benchmarks, ALFWorld and TextCraft, the planners achieve similar observed success in plan generation, while diffusion reduces mean plan-generation latency by 39-46% relative to AR. Our results show that Plan-and-Patch provides a framework for faster plan generation and effective plan repair in long-horizon agents.