AI Paper Atlas

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

Learning Predictive Models from Observation and Interaction

Karl Schmeckpeper, Annie Xie, Oleh Rybkin, Stephen Tian, Kostas Daniilidis, Sergey Levine, Chelsea Finn

ECCV 2020 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Embodied agent learning · 31 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 action-free BDD100K driving video via POISAVP trained on Boston actions · nuScenes Singapore action-conditioned prediction (driving)PSNR: 19.74 -> 20.88Improved
add action-free Singapore driving video via POISAVP trained on Boston actions · nuScenes Singapore action-conditioned prediction (driving)PSNR: 19.74 -> 20.81Improved
add action-free Singapore video via POISAVP trained on Boston actions · nuScenes Singapore action-conditioned prediction (driving)LPIPS: 0.1951 -> 0.1933Improved
add action-free BDD100K video via POISAVP trained on Boston actions · nuScenes Singapore action-conditioned prediction (driving)LPIPS: 0.1951 -> 0.2106Hurt
use POI instead of shared inverse modelaction-free Singapore video · nuScenes Singapore action-conditioned prediction (driving)PSNR: 20.65 -> 20.81No clear effect
use POI instead of CLASPaction-free Singapore video · nuScenes Singapore action-conditioned prediction (driving)PSNR: 20.57 -> 20.81No clear effect
add human tool-use video via POISAVP trained on random robot data · Robotic manipulation action-conditioned prediction (kinesthetic test set)PSNR: 23.31 -> 23.79Improved
add human tool-use video via POISAVP trained on random robot data · Robotic manipulation action-conditioned prediction (kinesthetic test set)SSIM: 0.803 -> 0.813Improved
add human tool-use video via POISAVP trained on random robot data · Robotic manipulation action-conditioned prediction (kinesthetic test set)LPIPS: 0.0757 -> 0.0722Improved
use POI with domain-dependent prior instead of shared inverse modelhuman video plus random robot data · Robotic manipulation action-conditioned prediction (kinesthetic test set)PSNR: 23.59 -> 23.79Improved
use CLASP instead of SAVP when adding human videorandom robot data · Robotic manipulation action-conditioned prediction (kinesthetic test set)PSNR: 23.31 -> 22.14Hurt
add human tool-use video via POISAVP in visual MPC · Robotic tool-use control (30 tasks, visual MPC)success rate (%): 23.3% -> 40.0%Improved

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