FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2996274873> ?p ?o ?g. }

• W2996274873 abstract "As deep reinforcement learning driven by visual perception becomes more widely used there is a growing need to better understand and probe the learned agents. Understanding the decision making process and its relationship to visual inputs can be very valuable to identify problems in learned behavior. However, this topic has been relatively under-explored in the research community. In this work we present a method for synthesizing visual inputs of interest for a trained agent. Such inputs or states could be situations in which specific actions are necessary. Further, critical states in which a very high or a very low reward can be achieved are often interesting to understand the situational awareness of the system as they can correspond to risky states. To this end, we learn a generative model over the state space of the environment and use its latent space to optimize a target function for the state of interest. In our experiments we show that this method can generate insights for a variety of environments and reinforcement learning methods. We explore results in the standard Atari benchmark games as well as in an autonomous driving simulator. Based on the efficiency with which we have been able to identify behavioural weaknesses with this technique, we believe this general approach could serve as an important tool for AI safety applications." @default.
• W2996274873 created "2019-12-26" @default.
• W2996274873 creator A5033763501 @default.
• W2996274873 creator A5075885606 @default.
• W2996274873 creator A5083153177 @default.
• W2996274873 date "2020-04-30" @default.
• W2996274873 modified "2023-09-26" @default.
• W2996274873 title "Finding and Visualizing Weaknesses of Deep Reinforcement Learning Agents" @default.
• W2996274873 cites W1514535095 @default.
• W2996274873 cites W1732154880 @default.
• W2996274873 cites W1836465849 @default.
• W2996274873 cites W1849277567 @default.
• W2996274873 cites W1903029394 @default.
• W2996274873 cites W1915485278 @default.
• W2996274873 cites W1932198206 @default.
• W2996274873 cites W1959608418 @default.
• W2996274873 cites W2112796928 @default.
• W2996274873 cites W2137278143 @default.
• W2996274873 cites W2163605009 @default.
• W2996274873 cites W2173564293 @default.
• W2996274873 cites W2187089797 @default.
• W2996274873 cites W2190008860 @default.
• W2996274873 cites W2253993278 @default.
• W2996274873 cites W2270696664 @default.
• W2996274873 cites W2295107390 @default.
• W2996274873 cites W2474163600 @default.
• W2996274873 cites W2557449848 @default.
• W2996274873 cites W2590082389 @default.
• W2996274873 cites W2594633041 @default.
• W2996274873 cites W2613718673 @default.
• W2996274873 cites W2616209507 @default.
• W2996274873 cites W2766047647 @default.
• W2996274873 cites W2798810981 @default.
• W2996274873 cites W2810348861 @default.
• W2996274873 cites W2891612330 @default.
• W2996274873 cites W2895261933 @default.
• W2996274873 cites W2897798332 @default.
• W2996274873 cites W2903075639 @default.
• W2996274873 cites W2945924974 @default.
• W2996274873 cites W2962851944 @default.
• W2996274873 cites W2962858109 @default.
• W2996274873 cites W2962954781 @default.
• W2996274873 cites W2963313316 @default.
• W2996274873 cites W2963464195 @default.
• W2996274873 cites W2963715038 @default.
• W2996274873 cites W2964043796 @default.
• W2996274873 cites W2964121744 @default.
• W2996274873 cites W3101609372 @default.
• W2996274873 hasPublicationYear "2020" @default.
• W2996274873 type Work @default.
• W2996274873 sameAs 2996274873 @default.
• W2996274873 citedByCount "1" @default.
• W2996274873 countsByYear W29962748732020 @default.
• W2996274873 crossrefType "proceedings-article" @default.
• W2996274873 hasAuthorship W2996274873A5033763501 @default.
• W2996274873 hasAuthorship W2996274873A5075885606 @default.
• W2996274873 hasAuthorship W2996274873A5083153177 @default.
• W2996274873 hasConcept C105795698 @default.
• W2996274873 hasConcept C107457646 @default.
• W2996274873 hasConcept C111472728 @default.
• W2996274873 hasConcept C111919701 @default.
• W2996274873 hasConcept C119857082 @default.
• W2996274873 hasConcept C13280743 @default.
• W2996274873 hasConcept C136197465 @default.
• W2996274873 hasConcept C138885662 @default.
• W2996274873 hasConcept C14036430 @default.
• W2996274873 hasConcept C154945302 @default.
• W2996274873 hasConcept C169760540 @default.
• W2996274873 hasConcept C17744445 @default.
• W2996274873 hasConcept C185798385 @default.
• W2996274873 hasConcept C199539241 @default.
• W2996274873 hasConcept C205649164 @default.
• W2996274873 hasConcept C26760741 @default.
• W2996274873 hasConcept C2778572836 @default.
• W2996274873 hasConcept C33923547 @default.
• W2996274873 hasConcept C39890363 @default.
• W2996274873 hasConcept C41008148 @default.
• W2996274873 hasConcept C63882131 @default.
• W2996274873 hasConcept C72434380 @default.
• W2996274873 hasConcept C78458016 @default.
• W2996274873 hasConcept C86803240 @default.
• W2996274873 hasConcept C9114305 @default.
• W2996274873 hasConcept C97541855 @default.
• W2996274873 hasConcept C98045186 @default.
• W2996274873 hasConceptScore W2996274873C105795698 @default.
• W2996274873 hasConceptScore W2996274873C107457646 @default.
• W2996274873 hasConceptScore W2996274873C111472728 @default.
• W2996274873 hasConceptScore W2996274873C111919701 @default.
• W2996274873 hasConceptScore W2996274873C119857082 @default.
• W2996274873 hasConceptScore W2996274873C13280743 @default.
• W2996274873 hasConceptScore W2996274873C136197465 @default.
• W2996274873 hasConceptScore W2996274873C138885662 @default.
• W2996274873 hasConceptScore W2996274873C14036430 @default.
• W2996274873 hasConceptScore W2996274873C154945302 @default.
• W2996274873 hasConceptScore W2996274873C169760540 @default.
• W2996274873 hasConceptScore W2996274873C17744445 @default.
• W2996274873 hasConceptScore W2996274873C185798385 @default.
• W2996274873 hasConceptScore W2996274873C199539241 @default.
• W2996274873 hasConceptScore W2996274873C205649164 @default.
• W2996274873 hasConceptScore W2996274873C26760741 @default.

• next