Matches in SemOpenAlex for { <https://semopenalex.org/work/W2913107815> ?p ?o ?g. }
- W2913107815 abstract "Complex environments and tasks pose a difficult problem for holistic end-to-end learning approaches. Decomposition of an environment into interacting controllable and non-controllable objects allows supervised learning for non-controllable objects and universal value function approximator learning for controllable objects. Such decomposition should lead to a shorter learning time and better generalisation capability. Here, we consider arcade-game environments as sets of interacting objects (controllable, non-controllable) and propose a set of functional modules that are specialized on mastering different types of interactions in a broad range of environments. The modules utilize regression, supervised learning, and reinforcement learning algorithms. Results of this case study in different Atari games suggest that human-level performance can be achieved by a learning agent within a human amount of game experience (10-15 minutes game time) when a proper decomposition of an environment or a task is provided. However, automatization of such decomposition remains a challenging problem. This case study shows how a model of a causal structure underlying an environment or a task can benefit learning time and generalization capability of the agent, and argues in favor of exploiting modular structure in contrast to using pure end-to-end learning approaches." @default.
- W2913107815 created "2019-02-21" @default.
- W2913107815 creator A5011185444 @default.
- W2913107815 creator A5065866182 @default.
- W2913107815 creator A5069344506 @default.
- W2913107815 creator A5074989840 @default.
- W2913107815 date "2019-01-27" @default.
- W2913107815 modified "2023-09-27" @default.
- W2913107815 title "Modularization of End-to-End Learning: Case Study in Arcade Games." @default.
- W2913107815 cites W1517008622 @default.
- W2913107815 cites W2145339207 @default.
- W2913107815 cites W2166822816 @default.
- W2913107815 cites W2613603362 @default.
- W2913107815 cites W2624780181 @default.
- W2913107815 cites W2733961795 @default.
- W2913107815 cites W2738724892 @default.
- W2913107815 cites W2761873684 @default.
- W2913107815 cites W2795776994 @default.
- W2913107815 cites W2888185177 @default.
- W2913107815 cites W2900717720 @default.
- W2913107815 cites W2903697572 @default.
- W2913107815 cites W2952115723 @default.
- W2913107815 cites W2963262099 @default.
- W2913107815 cites W2964001908 @default.
- W2913107815 cites W567721252 @default.
- W2913107815 hasPublicationYear "2019" @default.
- W2913107815 type Work @default.
- W2913107815 sameAs 2913107815 @default.
- W2913107815 citedByCount "4" @default.
- W2913107815 countsByYear W29131078152019 @default.
- W2913107815 countsByYear W29131078152020 @default.
- W2913107815 countsByYear W29131078152021 @default.
- W2913107815 crossrefType "posted-content" @default.
- W2913107815 hasAuthorship W2913107815A5011185444 @default.
- W2913107815 hasAuthorship W2913107815A5065866182 @default.
- W2913107815 hasAuthorship W2913107815A5069344506 @default.
- W2913107815 hasAuthorship W2913107815A5074989840 @default.
- W2913107815 hasConcept C101468663 @default.
- W2913107815 hasConcept C107457646 @default.
- W2913107815 hasConcept C119857082 @default.
- W2913107815 hasConcept C124681953 @default.
- W2913107815 hasConcept C127413603 @default.
- W2913107815 hasConcept C134306372 @default.
- W2913107815 hasConcept C14036430 @default.
- W2913107815 hasConcept C154945302 @default.
- W2913107815 hasConcept C177148314 @default.
- W2913107815 hasConcept C177264268 @default.
- W2913107815 hasConcept C18903297 @default.
- W2913107815 hasConcept C199360897 @default.
- W2913107815 hasConcept C201995342 @default.
- W2913107815 hasConcept C2780451532 @default.
- W2913107815 hasConcept C33923547 @default.
- W2913107815 hasConcept C41008148 @default.
- W2913107815 hasConcept C78458016 @default.
- W2913107815 hasConcept C86803240 @default.
- W2913107815 hasConcept C88482812 @default.
- W2913107815 hasConcept C97541855 @default.
- W2913107815 hasConceptScore W2913107815C101468663 @default.
- W2913107815 hasConceptScore W2913107815C107457646 @default.
- W2913107815 hasConceptScore W2913107815C119857082 @default.
- W2913107815 hasConceptScore W2913107815C124681953 @default.
- W2913107815 hasConceptScore W2913107815C127413603 @default.
- W2913107815 hasConceptScore W2913107815C134306372 @default.
- W2913107815 hasConceptScore W2913107815C14036430 @default.
- W2913107815 hasConceptScore W2913107815C154945302 @default.
- W2913107815 hasConceptScore W2913107815C177148314 @default.
- W2913107815 hasConceptScore W2913107815C177264268 @default.
- W2913107815 hasConceptScore W2913107815C18903297 @default.
- W2913107815 hasConceptScore W2913107815C199360897 @default.
- W2913107815 hasConceptScore W2913107815C201995342 @default.
- W2913107815 hasConceptScore W2913107815C2780451532 @default.
- W2913107815 hasConceptScore W2913107815C33923547 @default.
- W2913107815 hasConceptScore W2913107815C41008148 @default.
- W2913107815 hasConceptScore W2913107815C78458016 @default.
- W2913107815 hasConceptScore W2913107815C86803240 @default.
- W2913107815 hasConceptScore W2913107815C88482812 @default.
- W2913107815 hasConceptScore W2913107815C97541855 @default.
- W2913107815 hasLocation W29131078151 @default.
- W2913107815 hasOpenAccess W2913107815 @default.
- W2913107815 hasPrimaryLocation W29131078151 @default.
- W2913107815 hasRelatedWork W1595483645 @default.
- W2913107815 hasRelatedWork W2174786457 @default.
- W2913107815 hasRelatedWork W2283777287 @default.
- W2913107815 hasRelatedWork W2294805292 @default.
- W2913107815 hasRelatedWork W2611191801 @default.
- W2913107815 hasRelatedWork W2741926431 @default.
- W2913107815 hasRelatedWork W2757789407 @default.
- W2913107815 hasRelatedWork W2786916331 @default.
- W2913107815 hasRelatedWork W2789102901 @default.
- W2913107815 hasRelatedWork W2791039059 @default.
- W2913107815 hasRelatedWork W2920329255 @default.
- W2913107815 hasRelatedWork W2936382365 @default.
- W2913107815 hasRelatedWork W2948199691 @default.
- W2913107815 hasRelatedWork W2964263543 @default.
- W2913107815 hasRelatedWork W298069310 @default.
- W2913107815 hasRelatedWork W3014232838 @default.
- W2913107815 hasRelatedWork W3046387304 @default.
- W2913107815 hasRelatedWork W3126596751 @default.
- W2913107815 hasRelatedWork W3169104110 @default.
- W2913107815 hasRelatedWork W3206104157 @default.