AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Fair Policy Aggregation from Standard Policy Optimization

Ezgi Korkmaz

ICLR 2026 · Reinforcement learning › RL from feedback · approach: Reward modeling versus direct optimization

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
sample from uniform distribution on K* via policy-optimization oracle instead of uniform sampling over occupancy polytope Omax-quantile fairness · Factory monitoring MDP (10 agents, 5 warehouses)Nash welfare: 0.654 -> 0.701Improved
same change as abovemax-quantile fairness · Factory monitoring MDP (10 agents, 5 warehouses)Gini coefficient: 0.066 -> 0.087Hurt
same change as abovemax-quantile fairness · Factory monitoring MDP (10 agents, 5 warehouses)Max quantile q*: 0.999 -> 0.883Mixed
same change as abovemax-quantile fairness · Factory monitoring MDP (10 agents, 5 warehouses)wall-clock time: approximately 2 hours -> approximately 10 minutesImproved
use max-quantile fairness instead of utilitarian objectivepolicy aggregation · Factory monitoring MDP (10 agents, 5 warehouses)Nash welfare: 0.5736 -> 0.701Improved
use max-quantile fairness instead of egalitarian objectivepolicy aggregation · Factory monitoring MDP (10 agents, 5 warehouses)Gini coefficient: 0.2126 -> 0.087Improved

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