AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies

Hikmet Simsir, Ozgur S. Oguz

ICML 2026 · Reinforcement learning › Visual and representation RL · approach: Skill learning and diffusion policies · 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
replace unconstrained noise-space steering (DSRL) with Lagrangian trust-region residual perturbation (LP-DS)frozen diffusion/flow policies · Gym Walker2Depisodic return: roughly 4000 -> approximately 5000Improved
apply LP-DS to real Franka pick-and-place after simulation adaptationfrozen flow-matching backbone · Real Franka spatial pick-and-placesuccessful trials: 18/40 -> 33/40Improved
apply LP-DS to real Franka mug hangingfrozen flow-matching backbone · Real Franka mug hangingsuccessful trials: 11/20 -> 17/20Improved
add Lagrangian penalty on perturbation magnitude (ablation)LP-DS · Adroit Pensuccess rateImproved
add explicit hard bound on steering noise (ablation)LP-DS · Adroit Pensuccess rateNo clear effect
add both Lagrangian penalty and noise bound (ablation)LP-DS · Adroit Pensuccess rateImproved
add Lagrangian penalty to preserve action entropy (ablation)LP-DS · Adroit Penaction entropyImproved
apply residual perturbation in action space (LP-DS-A) instead of noise space (ablation)LP-DS · Walker2DreturnHurt
increase trust-region bound delta from 0.01/0.05 to 0.1 (ablation)LP-DS · toy multi-goal navigationmode coverageMixed
vary trust-region target delta (0.35, 0.5, 0.66) (ablation)LP-DS · Hopper, Walker2D, Robomimic Square, Adroit Relocatereward or successNo clear effect
use flow-matching backbone instead of diffusion backbone (ablation)LP-DS · Hopper-V2returnNo clear effect
use LP-DS instead of DSRL for entropy preservationnoise-space steering · Adroit Penk-NN action entropyImproved

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