Matches in SemOpenAlex for { <https://semopenalex.org/work/W3169239093> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W3169239093 endingPage "9858" @default.
- W3169239093 startingPage "9848" @default.
- W3169239093 abstract "Humans show an innate ability to learn the regularities of the world through interaction. By performing experiments in our environment, we are able to discern the causal factors of variation and infer how they affect the dynamics of our world. Analogously, here we attempt to equip reinforcement learning agents with the ability to perform experiments that facilitate a categorization of the rolled-out trajectories, and to subsequently infer the causal factors of the environment in a hierarchical manner. We introduce a novel intrinsic reward, called causal curiosity, and show that it allows our agents to learn optimal sequences of actions, and to discover causal factors in the dynamics. The learned behavior allows the agent to infer a binary quantized representation for the ground-truth causal factors in every environment. Additionally, we find that these experimental behaviors are semantically meaningful (e.g., to differentiate between heavy and light blocks, our agents learn to lift them), and are learnt in a self-supervised manner with approximately 2.5 times less data than conventional supervised planners. We show that these behaviors can be re-purposed and fine-tuned (e.g., from lifting to pushing or other downstream tasks). Finally, we show that the knowledge of causal factor representations aids zero-shot learning for more complex tasks." @default.
- W3169239093 created "2021-06-22" @default.
- W3169239093 creator A5014522247 @default.
- W3169239093 creator A5017710939 @default.
- W3169239093 creator A5044005697 @default.
- W3169239093 creator A5054494771 @default.
- W3169239093 creator A5064498575 @default.
- W3169239093 date "2021-05-04" @default.
- W3169239093 modified "2023-10-14" @default.
- W3169239093 title "Causal Curiosity: RL Agents Discovering Self-supervised Experiments for Causal Representation Learning" @default.
- W3169239093 hasPublicationYear "2021" @default.
- W3169239093 type Work @default.
- W3169239093 sameAs 3169239093 @default.
- W3169239093 citedByCount "2" @default.
- W3169239093 countsByYear W31692390932021 @default.
- W3169239093 crossrefType "proceedings-article" @default.
- W3169239093 hasAuthorship W3169239093A5014522247 @default.
- W3169239093 hasAuthorship W3169239093A5017710939 @default.
- W3169239093 hasAuthorship W3169239093A5044005697 @default.
- W3169239093 hasAuthorship W3169239093A5054494771 @default.
- W3169239093 hasAuthorship W3169239093A5064498575 @default.
- W3169239093 hasConcept C105795698 @default.
- W3169239093 hasConcept C11671645 @default.
- W3169239093 hasConcept C119857082 @default.
- W3169239093 hasConcept C121332964 @default.
- W3169239093 hasConcept C139002025 @default.
- W3169239093 hasConcept C154945302 @default.
- W3169239093 hasConcept C15744967 @default.
- W3169239093 hasConcept C163504300 @default.
- W3169239093 hasConcept C17744445 @default.
- W3169239093 hasConcept C199539241 @default.
- W3169239093 hasConcept C202269582 @default.
- W3169239093 hasConcept C2776359362 @default.
- W3169239093 hasConcept C33435437 @default.
- W3169239093 hasConcept C33923547 @default.
- W3169239093 hasConcept C41008148 @default.
- W3169239093 hasConcept C54355233 @default.
- W3169239093 hasConcept C62520636 @default.
- W3169239093 hasConcept C77805123 @default.
- W3169239093 hasConcept C86803240 @default.
- W3169239093 hasConcept C94124525 @default.
- W3169239093 hasConcept C94625758 @default.
- W3169239093 hasConcept C97541855 @default.
- W3169239093 hasConceptScore W3169239093C105795698 @default.
- W3169239093 hasConceptScore W3169239093C11671645 @default.
- W3169239093 hasConceptScore W3169239093C119857082 @default.
- W3169239093 hasConceptScore W3169239093C121332964 @default.
- W3169239093 hasConceptScore W3169239093C139002025 @default.
- W3169239093 hasConceptScore W3169239093C154945302 @default.
- W3169239093 hasConceptScore W3169239093C15744967 @default.
- W3169239093 hasConceptScore W3169239093C163504300 @default.
- W3169239093 hasConceptScore W3169239093C17744445 @default.
- W3169239093 hasConceptScore W3169239093C199539241 @default.
- W3169239093 hasConceptScore W3169239093C202269582 @default.
- W3169239093 hasConceptScore W3169239093C2776359362 @default.
- W3169239093 hasConceptScore W3169239093C33435437 @default.
- W3169239093 hasConceptScore W3169239093C33923547 @default.
- W3169239093 hasConceptScore W3169239093C41008148 @default.
- W3169239093 hasConceptScore W3169239093C54355233 @default.
- W3169239093 hasConceptScore W3169239093C62520636 @default.
- W3169239093 hasConceptScore W3169239093C77805123 @default.
- W3169239093 hasConceptScore W3169239093C86803240 @default.
- W3169239093 hasConceptScore W3169239093C94124525 @default.
- W3169239093 hasConceptScore W3169239093C94625758 @default.
- W3169239093 hasConceptScore W3169239093C97541855 @default.
- W3169239093 hasLocation W31692390931 @default.
- W3169239093 hasOpenAccess W3169239093 @default.
- W3169239093 hasPrimaryLocation W31692390931 @default.
- W3169239093 hasRelatedWork W1567876833 @default.
- W3169239093 hasRelatedWork W1970334321 @default.
- W3169239093 hasRelatedWork W2068820241 @default.
- W3169239093 hasRelatedWork W2106870251 @default.
- W3169239093 hasRelatedWork W2154521990 @default.
- W3169239093 hasRelatedWork W2155965100 @default.
- W3169239093 hasRelatedWork W2763238706 @default.
- W3169239093 hasRelatedWork W2790734939 @default.
- W3169239093 hasRelatedWork W2909534157 @default.
- W3169239093 hasRelatedWork W2995679616 @default.
- W3169239093 hasRelatedWork W2998653237 @default.
- W3169239093 hasRelatedWork W3011001566 @default.
- W3169239093 hasRelatedWork W3035090613 @default.
- W3169239093 hasRelatedWork W3037185591 @default.
- W3169239093 hasRelatedWork W3096689366 @default.
- W3169239093 hasRelatedWork W3129939205 @default.
- W3169239093 hasRelatedWork W3170800700 @default.
- W3169239093 hasRelatedWork W3192960001 @default.
- W3169239093 hasRelatedWork W3202842060 @default.
- W3169239093 hasRelatedWork W3206188474 @default.
- W3169239093 isParatext "false" @default.
- W3169239093 isRetracted "false" @default.
- W3169239093 magId "3169239093" @default.
- W3169239093 workType "article" @default.