English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent research has shown that graph neural networks (GNNs) can learnpolicies for locomotion control that are as effective as a typical multi-layerperceptron (MLP), with superior transfer and multi-task performance (Wang etal., 2018; Huang et al., 2020). Results have so far been limited to training onsmall agents, with the performance of GNNs deteriorating rapidly as the numberof sensors and actuators grows. A key motivation for the use of GNNs in thesupervised learning setting is their applicability to large graphs, but thisbenefit has not yet been realised for locomotion control. We identify theweakness with a common GNN architecture that causes this poor scaling:overfitting in the MLPs within the network that encode, decode, and propagatemessages. To combat this, we introduce Snowflake, a GNN training method forhigh-dimensional continuous control that freezes parameters in parts of thenetwork that suffer from overfitting. Snowflake significantly boosts theperformance of GNNs for locomotion control on large agents, now matching theperformance of MLPs, and with superior transfer properties.

Snowflake: Scaling GNNs to High-Dimensional Continuous Control via Parameter Freezing