AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Pitfall of Optimism: Distributional Reinforcement Learning by Randomizing Risk Criterion

Taehyun Cho, Seungyub Han, Heesoo Lee, Kyungjae Lee, Jungwoo Lee

NeurIPS 2023 · Reinforcement learning › Policy gradient and actor-critic · approach: Distributional value learning · 1 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
replace optimism with a randomized risk criterion (p-DLTV vs DLTV) in the stochastic N-Chain (ablation)DLTV · N-Chain (8,10)total count of optimal action: 9997 -> 14344Improved
replace optimism with randomized criterion at high intrinsic uncertainty (ablation)DLTV · N-Chain (1,17)total count of optimal action: 7257 -> 13404Improved
use PQR randomized criterion instead of epsilon-greedyQR-DQN · N-Chain (1,17)total count of optimal action: 11619 -> 14354Improved
increase variance of the suboptimal reward (distance between the two Gaussians) (ablation)DLTV · N-Chaintotal count of optimal action: 9997 -> 7257Hurt
add perturbation exploration to QR-DQN on Atari (PQR) under no-opsQR-DQN-zoo · Atari 55mean human-normalized score: 559 -> 1121Improved
PQR versus IQN under no-opsIQN-zoo · Atari 55mean human-normalized score: 902 -> 1121Improved
PQR versus Rainbow under sticky actionsRAINBOW-dopamine · Atari 55mean human-normalized score: 965 -> 962No clear effect
PQR versus Rainbow under no-opsRAINBOW-zoo · Atari 55mean human-normalized score: 1160 -> 1121Hurt
PQR versus QUOTA and DLTVPQR · Atari 55average human-normalized score: 603.66 -> 1078.00Improved
ablate schedule: fixed-size ambiguity set (1/t^0) instead of decaying (ablation)PQR · Pong-v4scoreHurt
ablate: force optimistic output vector (OPT) instead of random (ablation)PQR · Pong-v4scoreHurt
ablate: use sqrt(log t / t) schedule of p-DLTV (ablation)PQR · Pong-v4scoreHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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