AI Paper Atlas

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

Randomized Prior Functions for Deep Reinforcement Learning

Ian Osband, John Aslanides, Albin Cassirer

NeurIPS 2018 · Reinforcement learning › Policy gradient and actor-critic · approach: Regularized scalable deep value learning · 104 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 a fixed randomized prior function to each bootstrapped ensemble memberBootDQN · Chain environmentsepisodes to learnImproved
replace additive prior functions with L2 regularization on initial weightsBootDQN · Chain environmentsepisodes to learnHurt
add an additive prior function (ablation)Bootstrapped DQN · Chain environmentslargest problem solved before the episode limitImproved
vary the L2 regularization scale (ablation)BSR · Chain environmentslargest problem solved before the episode limitNo clear effect
increase ensemble size (ablation)BSP · Chain environmentslargest problem solved before the episode limitImproved
use small prior scales (ablation)BSP · Modified cartpole swing-upcumulative rewardHurt
increase prior scale (ablation)BSP · Modified cartpole swing-upcumulative rewardMixed
add randomized prior networksBootstrapped DQN · Modified cartpole swing-upcumulative rewardImproved
add randomized prior functionsBootstrapped DQN · Montezuma's Revengegame scoreImproved
use dropout or NoisyNet for explorationDQN · Chain environmentsepisodes to learnHurt
use state-visit-count UCB bonuses instead of randomized value functionsDQN · Chain environmentsepisodes to learnHurt
train independent ensemble members without prior functionsneural ensemble · Unobserved regions in 1D regressionpredictive uncertaintyHurt

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