AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

PaCo: Parameter-Compositional Multi-task Reinforcement Learning

Lingfeng Sun, Haichao Zhang, Wei Xu, Masayoshi Tomizuka

NeurIPS 2022 · Reinforcement learning › Visual and representation RL · approach: In-context and adaptive RL · 7 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 w-reset on top of loss maskout (ablation)PaCo · MT10-randsuccess rate (%): 76.0 -> 85.4Improved
add loss maskout on top of vanilla (no stabilization) (ablation)PaCo · MT10-randsuccess rate (%): 71.6 -> 76.0Improved
add both stabilization schemes to vanilla PaCo (ablation)PaCo · MT10-randsuccess rate (%): 71.6 -> 85.4Improved
raise parameter set size K from 3 to 5 (ablation)PaCo · MT10-randsuccess rate (%): 74.4 -> 85.4Improved
raise parameter set size K from 5 to 8 (ablation)PaCo · MT10-randsuccess rate (%): 85.4 -> 81.0Hurt
set K=10 with one-hot initialization of w (ablation)PaCo · MT10-randsuccess rate (%): 85.4 -> 88.9Improved
apply compositional structure to actor only instead of output layer only (ablation)PaCo · MT10-randsuccess rate (%): 64.0 -> 73.3Improved
apply compositional structure to actor and critic instead of actor only (ablation)PaCo · MT10-randsuccess rate (%): 73.3 -> 85.4Improved
PaCo versus CARE with task metadataPaCo · MT10-randsuccess rate (%): 76.0 -> 85.4Improved
PaCo versus CAREPaCo · MT10-fixedsuccess rate (%): 86.6 -> 93.3Improved
PaCo with K=20 versus CAREPaCo · MT50-randsuccess rate (%): 50.8 -> 57.3Improved
PaCo versus Multi-Task SACPaCo · MT10-randsuccess rate (%): 62.9 -> 85.4Improved

Later papers that compared with this method

6 later accepted papers report PaCo as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.

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