AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Autonomous Play with Correspondence-Driven Trajectory Warping

William Liang, Sam Wang, Hung-Ju Wang, Osbert Bastani, Yecheng Jason Ma, Dinesh Jayaraman

ICLR 2026 · Reinforcement learning › Visual and representation RL · approach: Vision-based robot RL

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 keypoint-correspondence imitation policy (KALM) with trajectory warpingTether · Real-world Pineapple to Bowlsuccess rate: 0.0 -> 1.0Improved
increase source demos from 1 to 10 (ablation)Tether · Real-world Wiping with Clothsuccess rate: 0.1 -> 0.6Improved
increase source demos from 1 to 10 (ablation)Tether · Real-world Pineapple to Cupsuccess rate: 0.5 -> 0.9Improved
increase source demos from 1 to 10 (ablation)Tether · Real-world Inserting Coffeesuccess rate: 0.2 -> 0.4Improved
increase source demos from 5 to 10 (ablation)Tether · Real-world Opening Cabinetsuccess rate: 0.3 -> 0.6Improved
increase source demos from 1 to 10 on easy pick-and-place (ablation)Tether · Real-world Strawberry to Bowlsuccess rate: 0.9 -> 0.9No clear effect
use Tether trajectory warping instead of zero-shot pi0Tether vs pi0 · Real-world Bowl to Shelfsuccess rate: 0.0 -> 0.9Improved
finetune pi0 on play-generated demos instead of 10 human demospi0-FAST-DROID · Real-world Pineapple to Bowlsuccess rate: 0.0 -> 0.8Improved
increase play-generated finetuning data from 50 to all demos (ablation)pi0-FAST-DROID · Real-world Bowl to Shelfsuccess rate: 0.1 -> 0.4Improved
use Gemini Robotics-ER 1.5 alone instead of GPT-4.1 with correspondence-based verification for success evaluation (ablation)Tether play success evaluator · Autonomous play annotated attemptsprecision: 97.7 -> 98.4Improved
use Gemini Robotics-ER 1.5 instead of Gemini-2.5-Flash for task planning (ablation)Tether play planner · Autonomous play annotated attemptstask planning accuracy: 94.1 -> 95.2Improved
use Tether policy instead of Diffusion Policy trained on 141-202 human demos for play statesAutonomous play policy · Play-induced initial states (20 per task)success rateImproved

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