Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312900898> ?p ?o ?g. }
- W4312900898 abstract "Training a high-dimensional simulated agent with an under-specified reward function often leads the agent to learn physically infeasible strategies that are ineffective when deployed in the real world. To mitigate these unnatural behaviors, reinforcement learning practitioners often utilize complex reward functions that encourage physically plausible behaviors. However, a tedious labor-intensive tuning process is often required to create hand-designed rewards which might not easily generalize across platforms and tasks. We propose substituting complex reward functions with “style rewards” learned from a dataset of motion capture demonstrations. A learned style reward can be combined with an arbitrary task reward to train policies that perform tasks using naturalistic strategies. These natural strategies can also facilitate transfer to the real world. We build upon Adversarial Motion Priors - an approach from the computer graphics domain that encodes a style reward from a dataset of reference motions - to demonstrate that an adversarial approach to training policies can produce behaviors that transfer to a real quadrupedal robot without requiring complex reward functions. We also demonstrate that an effective style reward can be learned from a few seconds of motion capture data gathered from a German Shepherd and leads to energy-efficient locomotion strategies with natural gait transitions." @default.
- W4312900898 created "2023-01-05" @default.
- W4312900898 creator A5015524477 @default.
- W4312900898 creator A5015686392 @default.
- W4312900898 creator A5036386828 @default.
- W4312900898 creator A5049349154 @default.
- W4312900898 creator A5050342525 @default.
- W4312900898 creator A5070606710 @default.
- W4312900898 creator A5086928709 @default.
- W4312900898 date "2022-10-23" @default.
- W4312900898 modified "2023-10-16" @default.
- W4312900898 title "Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions" @default.
- W4312900898 cites W1494650646 @default.
- W4312900898 cites W1994634749 @default.
- W4312900898 cites W1999874108 @default.
- W4312900898 cites W2090373601 @default.
- W4312900898 cites W2109026728 @default.
- W4312900898 cites W2116039916 @default.
- W4312900898 cites W2117778675 @default.
- W4312900898 cites W2133229776 @default.
- W4312900898 cites W2133793830 @default.
- W4312900898 cites W2151003079 @default.
- W4312900898 cites W2158794687 @default.
- W4312900898 cites W2593414223 @default.
- W4312900898 cites W2605102758 @default.
- W4312900898 cites W2788030459 @default.
- W4312900898 cites W2796290181 @default.
- W4312900898 cites W2811441028 @default.
- W4312900898 cites W2894766094 @default.
- W4312900898 cites W2899718839 @default.
- W4312900898 cites W2958043083 @default.
- W4312900898 cites W2962887844 @default.
- W4312900898 cites W2963184939 @default.
- W4312900898 cites W2963669336 @default.
- W4312900898 cites W2984579799 @default.
- W4312900898 cites W2987886924 @default.
- W4312900898 cites W3022265613 @default.
- W4312900898 cites W3104876774 @default.
- W4312900898 cites W3105609823 @default.
- W4312900898 cites W3147968035 @default.
- W4312900898 cites W3175254947 @default.
- W4312900898 cites W3206574325 @default.
- W4312900898 cites W3206620955 @default.
- W4312900898 cites W3206762371 @default.
- W4312900898 cites W4205430897 @default.
- W4312900898 cites W4226143977 @default.
- W4312900898 cites W4240805545 @default.
- W4312900898 doi "https://doi.org/10.1109/iros47612.2022.9981973" @default.
- W4312900898 hasPublicationYear "2022" @default.
- W4312900898 type Work @default.
- W4312900898 citedByCount "10" @default.
- W4312900898 countsByYear W43129008982022 @default.
- W4312900898 countsByYear W43129008982023 @default.
- W4312900898 crossrefType "proceedings-article" @default.
- W4312900898 hasAuthorship W4312900898A5015524477 @default.
- W4312900898 hasAuthorship W4312900898A5015686392 @default.
- W4312900898 hasAuthorship W4312900898A5036386828 @default.
- W4312900898 hasAuthorship W4312900898A5049349154 @default.
- W4312900898 hasAuthorship W4312900898A5050342525 @default.
- W4312900898 hasAuthorship W4312900898A5070606710 @default.
- W4312900898 hasAuthorship W4312900898A5086928709 @default.
- W4312900898 hasBestOaLocation W43129008982 @default.
- W4312900898 hasConcept C104114177 @default.
- W4312900898 hasConcept C107457646 @default.
- W4312900898 hasConcept C107673813 @default.
- W4312900898 hasConcept C111919701 @default.
- W4312900898 hasConcept C119857082 @default.
- W4312900898 hasConcept C127413603 @default.
- W4312900898 hasConcept C134306372 @default.
- W4312900898 hasConcept C154945302 @default.
- W4312900898 hasConcept C177769412 @default.
- W4312900898 hasConcept C201995342 @default.
- W4312900898 hasConcept C2780451532 @default.
- W4312900898 hasConcept C33923547 @default.
- W4312900898 hasConcept C36503486 @default.
- W4312900898 hasConcept C37736160 @default.
- W4312900898 hasConcept C41008148 @default.
- W4312900898 hasConcept C97541855 @default.
- W4312900898 hasConcept C98045186 @default.
- W4312900898 hasConceptScore W4312900898C104114177 @default.
- W4312900898 hasConceptScore W4312900898C107457646 @default.
- W4312900898 hasConceptScore W4312900898C107673813 @default.
- W4312900898 hasConceptScore W4312900898C111919701 @default.
- W4312900898 hasConceptScore W4312900898C119857082 @default.
- W4312900898 hasConceptScore W4312900898C127413603 @default.
- W4312900898 hasConceptScore W4312900898C134306372 @default.
- W4312900898 hasConceptScore W4312900898C154945302 @default.
- W4312900898 hasConceptScore W4312900898C177769412 @default.
- W4312900898 hasConceptScore W4312900898C201995342 @default.
- W4312900898 hasConceptScore W4312900898C2780451532 @default.
- W4312900898 hasConceptScore W4312900898C33923547 @default.
- W4312900898 hasConceptScore W4312900898C36503486 @default.
- W4312900898 hasConceptScore W4312900898C37736160 @default.
- W4312900898 hasConceptScore W4312900898C41008148 @default.
- W4312900898 hasConceptScore W4312900898C97541855 @default.
- W4312900898 hasConceptScore W4312900898C98045186 @default.
- W4312900898 hasLocation W43129008981 @default.
- W4312900898 hasLocation W43129008982 @default.
- W4312900898 hasOpenAccess W4312900898 @default.
- W4312900898 hasPrimaryLocation W43129008981 @default.