Matches in SemOpenAlex for { <https://semopenalex.org/work/W4382318533> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4382318533 endingPage "16177" @default.
- W4382318533 startingPage "16176" @default.
- W4382318533 abstract "It has been shown that an agent can be trained with an adversarial policy which achieves high degrees of success against a state-of-the-art DRL victim despite taking unintuitive actions. This prompts the question: is this adversarial behaviour detectable through the observations of the victim alone? We find that widely used classification methods such as random forests are only able to achieve a maximum of ≈71% test set accuracy when classifying an agent for a single timestep. However, when the classifier inputs are treated as time-series data, test set classification accuracy is increased significantly to ≈98%. This is true for both classification of episodes as a whole, and for “live” classification at each timestep in an episode. These classifications can then be used to “react” to incoming attacks and increase the overall win rate against Adversarial opponents by approximately 17%. Classification of the victim’s own internal activations in response to the adversary is shown to achieve similarly impressive accuracy while also offering advantages like increased transferability to other domains." @default.
- W4382318533 created "2023-06-28" @default.
- W4382318533 creator A5046330057 @default.
- W4382318533 creator A5088726212 @default.
- W4382318533 creator A5092346213 @default.
- W4382318533 date "2023-06-26" @default.
- W4382318533 modified "2023-09-30" @default.
- W4382318533 title "Know Your Enemy: Identifying Adversarial Behaviours in Deep Reinforcement Learning Agents (Student Abstract)" @default.
- W4382318533 doi "https://doi.org/10.1609/aaai.v37i13.26948" @default.
- W4382318533 hasPublicationYear "2023" @default.
- W4382318533 type Work @default.
- W4382318533 citedByCount "0" @default.
- W4382318533 crossrefType "journal-article" @default.
- W4382318533 hasAuthorship W4382318533A5046330057 @default.
- W4382318533 hasAuthorship W4382318533A5088726212 @default.
- W4382318533 hasAuthorship W4382318533A5092346213 @default.
- W4382318533 hasBestOaLocation W43823185331 @default.
- W4382318533 hasConcept C119857082 @default.
- W4382318533 hasConcept C140331021 @default.
- W4382318533 hasConcept C154945302 @default.
- W4382318533 hasConcept C169258074 @default.
- W4382318533 hasConcept C177264268 @default.
- W4382318533 hasConcept C199360897 @default.
- W4382318533 hasConcept C37736160 @default.
- W4382318533 hasConcept C38652104 @default.
- W4382318533 hasConcept C41008148 @default.
- W4382318533 hasConcept C41065033 @default.
- W4382318533 hasConcept C61272859 @default.
- W4382318533 hasConcept C95623464 @default.
- W4382318533 hasConcept C97541855 @default.
- W4382318533 hasConceptScore W4382318533C119857082 @default.
- W4382318533 hasConceptScore W4382318533C140331021 @default.
- W4382318533 hasConceptScore W4382318533C154945302 @default.
- W4382318533 hasConceptScore W4382318533C169258074 @default.
- W4382318533 hasConceptScore W4382318533C177264268 @default.
- W4382318533 hasConceptScore W4382318533C199360897 @default.
- W4382318533 hasConceptScore W4382318533C37736160 @default.
- W4382318533 hasConceptScore W4382318533C38652104 @default.
- W4382318533 hasConceptScore W4382318533C41008148 @default.
- W4382318533 hasConceptScore W4382318533C41065033 @default.
- W4382318533 hasConceptScore W4382318533C61272859 @default.
- W4382318533 hasConceptScore W4382318533C95623464 @default.
- W4382318533 hasConceptScore W4382318533C97541855 @default.
- W4382318533 hasIssue "13" @default.
- W4382318533 hasLocation W43823185331 @default.
- W4382318533 hasOpenAccess W4382318533 @default.
- W4382318533 hasPrimaryLocation W43823185331 @default.
- W4382318533 hasRelatedWork W2460937040 @default.
- W4382318533 hasRelatedWork W2913608505 @default.
- W4382318533 hasRelatedWork W2946919881 @default.
- W4382318533 hasRelatedWork W2952541330 @default.
- W4382318533 hasRelatedWork W2953083558 @default.
- W4382318533 hasRelatedWork W3037567525 @default.
- W4382318533 hasRelatedWork W3037770290 @default.
- W4382318533 hasRelatedWork W4249229055 @default.
- W4382318533 hasRelatedWork W4300511536 @default.
- W4382318533 hasRelatedWork W4382318533 @default.
- W4382318533 hasVolume "37" @default.
- W4382318533 isParatext "false" @default.
- W4382318533 isRetracted "false" @default.
- W4382318533 workType "article" @default.