[KongchangAI]
PaperPlan-and-Patch: Diffusion Language Models for Agentic Planning

Plan-and-Patch

A research paper proposing the use of diffusion language models for agent planning with local repair. The core idea is that when a plan requires revision, only the affected regions are locally regenerated (patched) rather than the entire plan being discarded and rewritten from scratch.

Timeline (last 90 days)

Oct 9

The arXiv paper "Plan-and-Patch: Diffusion Language Models for Agentic Planning" proposes using diffusion language models to achieve agentic planning and local repair.

Unverified50%
Oct 9

Plan-and-Patch is a plan-and-act framework whose core is using a diffusion language model (dLLM) to generate and repair plans.

Unverified50%
Oct 9

Based on a masked-filling mechanism, diffusion models can retain prefixes and suffixes and conditionally refill excised middle segments, enabling local repair.

Unverified50%

All Facts (3)

Source Articles