AI Paper Atlas

Research map / Reinforcement learning / MDPs and planning

The Impact of Task Underspecification in Evaluating Deep Reinforcement Learning

Vindula Jayawardana, Catherine H Tang, Sirui Li, Dajiang Suo, Cathy Wu

NeurIPS 2022 · Reinforcement learning › MDPs and planning · approach: POMDPs and confounded RL · 4 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
Evaluate traffic signal controllers across an MDP family instead of a single intersectionTraffic signal control methods · Traffic signal controlrelative method ranking: IPPO: 0.84 -> 1.7Mixed
Evaluate methods across a task's MDP family instead of its standard point MDPDRL methods · Cartpolerelative method ranking: PPO 61%; TRPO 56%Mixed
Use k-means clustering to select representative point MDPsMDP-family performance estimation · Traffic signal control and control-task familiesapproximation accuracyImproved
Use random sampling with replacements rather than representative clusteringMDP-family performance estimation · Control-task familiesapproximation varianceHurt
Use random sampling without replacements rather than representative clusteringMDP-family performance estimation · Traffic signal controlapproximation accuracyMixed
Use random sampling without replacements rather than sampling with replacementsMDP-family performance estimation · Cartpole, Pendulum, and Half-cheetahapproximation accuracyImproved
Evaluate methods on a simplified point MDP rather than other family instancesDRL methods · Quad, Pendulum, and Swimmertask performanceMixed
Evaluate on one selected point MDP rather than a family of MDPsDRL methods · Underspecified tasksreliability of relative rankingsHurt

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