AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

RoboHive: A Unified Framework for Robot Learning

Vikash Kumar, Rutav Shah, Gaoyue Zhou, Vincent Moens, Vittorio Caggiano, Abhishek Gupta, Aravind Rajeswaran

NeurIPS 2023 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Embodied agent learning · 3 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 visual observations to state observationsRoboHive environments · Arm Manipulationsamples per second: 7473 -> 993Hurt
add visual observations to state observationsRoboHive environments · Hand manipulationsamples per second: 8785 -> 1713Hurt
add visual observations to state observationsRoboHive environments · Multi-tasksamples per second: 4430 -> 346Hurt
use pixel (v2d) observations instead of stateFrankaReachRandom · Arm Manipulationsamples per second: 8911.4261 -> 1055.2824Hurt
use pixel (v2d) observations instead of staterelocate · Hand manipulationsamples per second: 9317.9963 -> 1737.4508Hurt
use visual representation baselines (RRL, R3M, VC1) instead of privileged state policybehavior cloning · Kitchen, Adroit and D'Kitty datasetstask performanceHurt
use visual representations instead of proprioception-only inputbehavior cloning · D'Kitty locomotion (ROBEL)task performanceHurt
train on human teleoperated data instead of expert policy datavisual baselines · Kitchen suitetask performanceHurt

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