AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

PACT: Self-Evolving Physical Safety Alignment for Diffusion Policies in Embodied Manipulation

Lingxuan Wu, Zijian Zhu, Lizhong Wang, Chengyang Ying, Huayu Chen, Xiao Yang, Fangming Liu, Jun Zhu

ICML 2026 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
post-train with PACTDP · RoboTwin 7 tasks avgsuccess rate: 45% -> 73%Improved
post-train with PACTDP · RoboTwin 7 tasks avgsafe rate: 41% -> 65%Improved
post-train with PACTDP3 · RoboTwin 7 tasks avgsuccess rate: 27% -> 43%Improved
post-train with PACTRDT-1B · RoboTwin 7 tasks avgsuccess rate: 56% -> 79%Improved
post-train with PACTpi0.5 · RoboTwin 7 tasks avgsafe rate: 37% -> 62%Improved
use PACT distillation instead of PPODP · RoboTwin Pour Watersafe rate: 31% -> 64%Improved
use PACT distillation instead of DIPODP · RoboTwin Pour Watersuccess rate: 67% -> 86%Improved
use on-policy PACT instead of off-policy distillation on guided rollouts (ablation)DP · RoboTwin Pour Watersuccess rate: 72% -> 86%Improved
use curriculum distillation instead of direct distillation (ablation)PACT · RoboTwintraining stabilityImproved
increase Lagrange multiplier lambda too large (ablation)PACT · RoboTwinperformanceHurt
inject cost gradient over larger max diffusion time t_c (ablation)PACT · RoboTwinperformanceHurt
raise UTD ratio from 100 to 200 (288 rollouts) (ablation)PACT · RoboTwin Handover Applesuccess rate: 75% -> 45%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy