AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Much Ado About Noising: Dispelling the Myths of Generative Robotic Control

Chaoyi Pan, Giri Anantharaman, Nai-Chieh Huang, Claire Jin, Daniel Pfrommer, Chenyang Yuan, Frank Permenter, Guannan Qu, Nicholas Matthew Boffi, Guanya Shi, Max Simchowitz

ICLR 2026 · Reinforcement learning › Imitation learning · approach: Imitation combined with RL · 36 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
add stochasticity injection and supervised iterative computation (MIP) to regression (ablation)regression policy · Tool-Hang (state)success rate: 0.54 -> 0.80Improved
add iterative computation alone (residual regression) (ablation)regression policy · Tool-Hang (state)success rate: Regression << RR 0.54Improved
increase integration steps of straight flow from 1 to 9 (ablation)straight flow · Tool-Hang (state)success rate: 0.54 -> 0.52No clear effect
supervise intermediate steps (ablation)MIP · Tool-Hang (state)success rate: 0.42 -> 0.80Improved
MIP vs flow with fewer integration steps (2 vs 9 NFEs)MIP · LIBERO 10success rate: 81.6 -> 82.2No clear effect
use l1 instead of l2 loss for regression (ablation)regression policy (Sudeep-DiT) · Tool-Hang ph (state)success rate: 0.50/0.37 -> 0.76/0.65Improved
use l1 instead of l2 loss for l1 on LIBERO regression (ablation)pi0 regression · LIBERO 10success rate: 78.0 -> 62.4Hurt
perturb observations during training (ablation)MIP · Transport / Tool-Hang (state)success rate: MIP 0.92/0.85 -> MIP-Obs 0.13/0.08 (Tool-Hang, Chi-Transformer)Hurt
use flow instead of regressionpolicy · Tool-Hang (state)policy Lipschitz constant: 1.71 -> 1.41Improved
MIP vs consistency trajectory model at 2 stepsMIP · Tool-Hang (state)success rate: 0.56/0.26 -> 0.92/0.88Improved
train on lower-quality (delayed and noisy) data, flow vs regressionChi-UNet · Tool-Hang data qualitysuccess rate: regression 0.70/0.63, flow 0.76/0.68No clear effect

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