AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

PACE: Parameter Change for Unsupervised Environment Design

Fang YUAN, Junjie Zeng, Qinglun Li, Long Qin, Quanjun Yin, Siqi Shen, Yuxiang Xie, Junqiang Yang

ICML 2026 · Reinforcement learning › Visual and representation RL · approach: Inferring decision-making from behavior · 0 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
score levels by squared parameter change of a provisional update instead of GAE or MaxMC proxiesPACE (PPO student) · MiniGrid (12 held-out levels)IQM of solved rate: 0.808 -> 0.964Improved
score levels by squared parameter change instead of proxy scoresPACE (PPO student) · MiniGrid (12 held-out levels)Optimality Gap: 0.299 -> 0.172Improved
score levels by squared parameter change instead of proxy scoresPACE (PPO student) · MiniGrid PerfectMaze (51x51)solve rate (%): 72.3% vs baselines 26.33-54.17%Improved
score levels by squared parameter change instead of proxy scoresPACE (PPO student) · Craftax-1BIQM of reward: 0.618 -> 0.694Improved
score levels by squared parameter change instead of proxy scoresPACE (PPO student) · Craftax-1Bevaluation reward: 31.198 ± 2.439Improved
use PACE scoring instead of PLR/ACCEL on easy MiniGrid levels (FourRooms, Crossing)PACE (PPO student) · MiniGrid FourRooms and Crossingsolved rate: all methods above 84% on FourRoomsNo clear effect
use PACE scoring instead of PLR⊥ on 16RoomsFewerDoorsPACE (PPO student) · MiniGrid 16RoomsFewerDoorssolved rate: 0.972 -> 0.728Hurt
use PACE scoring instead of PLR on large PerfectMaze(M) levelPACE (PPO student) · MiniGrid PerfectMaze(M)solved rate: 0.373 -> 0.359No clear effect
use parameter-change score instead of PLR/PLR⊥/ACCEL scoresPACE · MiniGrid score stabilityMean CV (lower better): 0.0055 (PLR⊥) -> 0.0010Improved
use parameter-change score instead of PLR/PLR⊥/ACCEL scoresPACE · Craftax score stabilityMean CV (lower better): 0.6105 (PLR) -> 0.5412Improved
regret-based curation (ACCEL) versus domain randomization in CraftaxACCEL · Craftax-1Bevaluation rewardNo clear effect

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