AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

On the Effectiveness of Fine-tuning Versus Meta-reinforcement Learning

Zhao Mandi, Pieter Abbeel, Stephen James

NeurIPS 2022 · Reinforcement learning › Visual and representation RL · approach: Task-inference meta-RL · 19 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
multi-task pretraining + fine-tuning instead of Reptile meta-pretrainingPPO on Procgen Coinrun · unseen Coinrun levelsreturn after adaptationImproved
RL2 test-time adaptation instead of fine-tuningPPO on Procgen Coinrun · unseen Coinrun levelsreturn after adaptationHurt
increase number of pretraining levels (100 to 10000) (ablation)MT-PPO · Procgen Coinrun unseen levelszero-shot and fine-tuned returnImproved
multi-task fine-tuning vs Reptile meta-pretrainingC2F-ARM · RLBench 5 unseen tasks, 0 demossuccess rateNo clear effect
multi-task fine-tuning vs PEARL context adaptationC2F-ARM · RLBench 5 unseen taskssuccess rateImproved
multi-task pretraining then fine-tuning instead of training from scratchC2F-ARM · RLBench unseen tasks, sparse rewardsuccess rateImproved
fine-tune unseen task in isolation instead of jointly with training tasks (ablation)MT-C2F-ARM · RLBench unseen taskssuccess rateImproved
fine-tune both Q-attention depths instead of only the fine depth (ablation)MT-C2F-ARM · RLBench unseen taskssuccess rateImproved
multi-task pretraining then fine-tuning vs training from scratchRainbow DQN · Atari 5 unseen gamesnormalized scoreNo clear effect
fine-tune only convolutional layers (reinit MLP) instead of all parameters (ablation)MT-Rainbow · Atari unseen gamesnormalized scoreNo clear effect
add test-time fine-tuning to PEARL-Rainbow (ablation)PEARL-Rainbow · Atari unseen gamesnormalized scoreImproved
PEARL gradient-free context adaptationC2F-ARM/Rainbow · RLBench and Atari unseen tasksadaptation performanceHurt

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