FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3134075242> ?p ?o ?g. }

• W3134075242 abstract "Recent research has shown that Graph Neural Networks (GNNs) can learn policies for locomotion control that are as effective as a typical multi-layer perceptron (MLP), with superior transfer and multi-task performance (Wang et al., 2018; Huang et al., 2020). Results have so far been limited to training on small agents, with the performance of GNNs deteriorating rapidly as the number of sensors and actuators grows. A key motivation for the use of GNNs in the supervised learning setting is their applicability to large graphs, but this benefit has not yet been realised for locomotion control. We identify the weakness with a common GNN architecture that causes this poor scaling: overfitting in the MLPs within the network that encode, decode, and propagate messages. To combat this, we introduce Snowflake, a GNN training method for high-dimensional continuous control that freezes parameters in parts of the network that suffer from overfitting. Snowflake significantly boosts the performance of GNNs for locomotion control on large agents, now matching the performance of MLPs, and with superior transfer properties." @default.
• W3134075242 created "2021-03-15" @default.
• W3134075242 creator A5024136382 @default.
• W3134075242 creator A5043731054 @default.
• W3134075242 creator A5056879203 @default.
• W3134075242 creator A5080745300 @default.
• W3134075242 date "2021-03-01" @default.
• W3134075242 modified "2023-09-27" @default.
• W3134075242 title "Snowflake: Scaling GNNs to High-Dimensional Continuous Control via Parameter Freezing." @default.
• W3134075242 cites W1480330138 @default.
• W3134075242 cites W1533861849 @default.
• W3134075242 cites W1614298861 @default.
• W3134075242 cites W1771410628 @default.
• W3134075242 cites W1845972764 @default.
• W3134075242 cites W2089497633 @default.
• W3134075242 cites W2095705004 @default.
• W3134075242 cites W2116341502 @default.
• W3134075242 cites W2130801532 @default.
• W3134075242 cites W2145339207 @default.
• W3134075242 cites W2149933564 @default.
• W3134075242 cites W2157331557 @default.
• W3134075242 cites W2158782408 @default.
• W3134075242 cites W2167729035 @default.
• W3134075242 cites W2250539671 @default.
• W3134075242 cites W2299115575 @default.
• W3134075242 cites W2606780347 @default.
• W3134075242 cites W2607264901 @default.
• W3134075242 cites W2726187156 @default.
• W3134075242 cites W2736601468 @default.
• W3134075242 cites W2785738552 @default.
• W3134075242 cites W2785948534 @default.
• W3134075242 cites W2805516822 @default.
• W3134075242 cites W2883638665 @default.
• W3134075242 cites W2907502844 @default.
• W3134075242 cites W2949608212 @default.
• W3134075242 cites W2950898568 @default.
• W3134075242 cites W2963184621 @default.
• W3134075242 cites W2963207497 @default.
• W3134075242 cites W2963403868 @default.
• W3134075242 cites W2963504252 @default.
• W3134075242 cites W2963771282 @default.
• W3134075242 cites W2963990127 @default.
• W3134075242 cites W2964121744 @default.
• W3134075242 cites W2964303773 @default.
• W3134075242 cites W2969996838 @default.
• W3134075242 cites W2970292059 @default.
• W3134075242 cites W2994892931 @default.
• W3134075242 cites W3007309629 @default.
• W3134075242 cites W3033100261 @default.
• W3134075242 cites W3034275286 @default.
• W3134075242 cites W3034847664 @default.
• W3134075242 cites W3034955736 @default.
• W3134075242 cites W3035407756 @default.
• W3134075242 cites W3036295212 @default.
• W3134075242 cites W3037207827 @default.
• W3134075242 cites W3087257704 @default.
• W3134075242 cites W3105390215 @default.
• W3134075242 cites W3106504817 @default.
• W3134075242 cites W3107340688 @default.
• W3134075242 cites W3115727409 @default.
• W3134075242 cites W3123302689 @default.
• W3134075242 cites W3181414820 @default.
• W3134075242 hasPublicationYear "2021" @default.
• W3134075242 type Work @default.
• W3134075242 sameAs 3134075242 @default.
• W3134075242 citedByCount "0" @default.
• W3134075242 crossrefType "posted-content" @default.
• W3134075242 hasAuthorship W3134075242A5024136382 @default.
• W3134075242 hasAuthorship W3134075242A5043731054 @default.
• W3134075242 hasAuthorship W3134075242A5056879203 @default.
• W3134075242 hasAuthorship W3134075242A5080745300 @default.
• W3134075242 hasConcept C119857082 @default.
• W3134075242 hasConcept C121332964 @default.
• W3134075242 hasConcept C153294291 @default.
• W3134075242 hasConcept C154945302 @default.
• W3134075242 hasConcept C162324750 @default.
• W3134075242 hasConcept C187736073 @default.
• W3134075242 hasConcept C197046000 @default.
• W3134075242 hasConcept C22019652 @default.
• W3134075242 hasConcept C2775924081 @default.
• W3134075242 hasConcept C2780451532 @default.
• W3134075242 hasConcept C41008148 @default.
• W3134075242 hasConcept C50644808 @default.
• W3134075242 hasConcept C54059038 @default.
• W3134075242 hasConcept C60908668 @default.
• W3134075242 hasConceptScore W3134075242C119857082 @default.
• W3134075242 hasConceptScore W3134075242C121332964 @default.
• W3134075242 hasConceptScore W3134075242C153294291 @default.
• W3134075242 hasConceptScore W3134075242C154945302 @default.
• W3134075242 hasConceptScore W3134075242C162324750 @default.
• W3134075242 hasConceptScore W3134075242C187736073 @default.
• W3134075242 hasConceptScore W3134075242C197046000 @default.
• W3134075242 hasConceptScore W3134075242C22019652 @default.
• W3134075242 hasConceptScore W3134075242C2775924081 @default.
• W3134075242 hasConceptScore W3134075242C2780451532 @default.
• W3134075242 hasConceptScore W3134075242C41008148 @default.
• W3134075242 hasConceptScore W3134075242C50644808 @default.
• W3134075242 hasConceptScore W3134075242C54059038 @default.
• W3134075242 hasConceptScore W3134075242C60908668 @default.
• W3134075242 hasLocation W31340752421 @default.

• next