Apple Proposes DACA-GRPO: Denoising-Aware Credit Assignment for Diffusion LLM RL
Apple ML Research · rss · 2026-09-16
Apple ML Research introduces DACA-GRPO (Denoising-Aware Credit Assignment for GRPO), targeting two fundamental weaknesses of RL for diffusion LLMs:
- Problem: existing methods treat all denoising steps as equally important, lacking temporal credit assignment across the denoising trajectory, and rely on biased, high-variance mean-field likelihood estimates
- Solution: DACA-GRPO is a lightweight, plug-and-play enhancement for any GRPO-style trainer, adding denoising-aware credit assignment
A method-level advance for RL fine-tuning of diffusion language models.
More from Research
- Radial gate objective cuts diffusion transformer FID from 16.24 to 11.83 — Junha Park · 2026-11-02
- REA: async brain-cerebellum runtime for robots where policies never wait on the LLM — Pengfei Zhou · 2026-11-02
- Microsoft's MemSteer turns unlabeled user histories into agent memory supervision — MicrosoftResearch · 2026-11-02
- SciGame benchmark uses counterfactual worlds to expose LLM agents' priors vs evidence — Yusheng Zhao · 2026-11-02
- On Terence Tao's blog, Frank Calegari calls on mathematicians to adapt after OpenAI's 700-paper release — Hesamation · 2026-11-02
- Tsinghua & Microsoft's HarnessMM cuts agent token use by up to 87.93% while raising success rate — rohanpaul_ai · 2026-11-02