AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Modified Frank Wolfe in Probability Space

Carson Kent, Jiajin Li, Jose Blanchet, Peter Glynn

NeurIPS 2021 · Optimization › Nonconvex and smooth optimization · approach: Stochastic Frank-Wolfe methods · 2 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 robust MCE IRL (opponent strength 1-alpha) on top of standard MCE IRLMCE IRL · GridWorld and ObjectWorld under dynamics mismatchvalue of learned policy in learner MDPImproved
use robust MCE IRL when expert noise is low (little mismatch)Robust MCE IRL · GridWorld and ObjectWorld, low expert noisevalue of learned policy in learner MDPHurt
underestimate the opponent strength 1-alpha so the expert dynamics lie outside the uncertainty set (ablation)Robust MCE IRL · GridWorld and ObjectWorldvalue of learned policy in learner MDPHurt
overestimate the opponent strength 1-alpha (ablation)Robust MCE IRL · GridWorld and ObjectWorldvalue of learned policy in learner MDPHurt
set 1-alpha near half the distance between learner dynamics and a rough estimate of expert dynamics (ablation)Robust MCE IRL · GridWorld and ObjectWorldvalue of learned policy in learner MDPImproved
increase dynamics mismatch (expert noise) for standard MCE IRLStandard MCE IRL · GridWorld and ObjectWorldvalue of learned policy in learner MDPHurt
add robust variant to Relative Entropy IRL in continuous state-action spaceRE IRL · GaussianGridvalue of learned policy in learner MDPImproved
add robust training to MCE IRL with low-dimensional feature mappingMCE IRL · low-dimensional feature GridWorld (appendix)value of learned policy in learner MDPImproved
choose alpha = 1 - d_dyn/2 in the constructive three-state exampleRobust MCE IRL · constructive example MDPperformance gap to no-mismatch MCE IRLImproved

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