AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Information Maximizing Curriculum: A Curriculum-Based Approach for Learning Versatile Skills

Denis Blessing, Onur Celik, Xiaogang Jia, Moritz Reuss, Maximilian Xiling Li, Rudolf Lioutikov, Gerhard Neumann

NeurIPS 2023 · Reinforcement learning › Imitation learning · approach: Behavior cloning theory and models

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
train mixture of experts with IMC instead of EMmixture of experts · Obstacle Avoidancesuccess rate: 0.675 -> 0.855Improved
train mixture of experts with IMC instead of EMmixture of experts · Obstacle Avoidanceentropy: 0.902 -> 0.930Improved
train mixture of experts with IMC instead of EMmixture of experts · Block Pushingsuccess rate: 0.458 -> 0.521Improved
train mixture of experts with IMC instead of EMmixture of experts · Block Pushingentropy: 0.707 -> 0.654Hurt
train mixture of experts with IMC instead of EMmixture of experts · Table Tennisdistance error: 0.220 -> 0.153Improved
train mixture of experts with IMC instead of EMmixture of experts · Table Tennissuccess rate: 0.725 -> 0.870Improved
use IMC instead of DDPMpolicy · Table Tennisdistance error: 0.185 -> 0.153Improved
use IMC instead of CVAEpolicy · Obstacle Avoidanceentropy: 0.465 -> 0.930Improved
use IMC instead of CVAEpolicy · Obstacle Avoidancesuccess rate: 0.853 -> 0.855No clear effect
use IMC instead of BETpolicy · Obstacle Avoidancesuccess rate: 0.504 -> 0.855Improved
use IMC instead of ML-Curpolicy · Block Pushingsuccess rate: 0.000 -> 0.521Improved
vary number of components for IMC vs EM (1 to 50), few components (ablation)IMC vs EM · Obstacle Avoidance / Table Tennissuccess 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