Matches in SemOpenAlex for { <https://semopenalex.org/work/W4311640020> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W4311640020 abstract "Our work targets at searching feasible adversarial perturbation to attack a classifier with high-dimensional categorical inputs in a domain-agnostic setting. This is intrinsically an NP-hard knapsack problem where the exploration space becomes explosively larger as the feature dimension increases. Without the help of domain knowledge, solving this problem via heuristic method, such as Branch-and-Bound, suffers from exponential complexity, yet can bring arbitrarily bad attack results. We address the challenge via the lens of multi-armed bandit based combinatorial search. Our proposed method, namely FEAT, treats modifying each categorical feature as pulling an arm in multi-armed bandit programming. Our objective is to achieve highly efficient and effective attack using an Orthogonal Matching Pursuit (OMP)-enhanced Upper Confidence Bound (UCB) exploration strategy. Our theoretical analysis bounding the regret gap of FEAT guarantees its practical attack performance. In empirical analysis, we compare FEAT with other state-of-the-art domain-agnostic attack methods over various real-world categorical data sets of different applications. Substantial experimental observations confirm the expected efficiency and attack effectiveness of FEAT applied in different application scenarios. Our work further hints the applicability of FEAT for assessing the adversarial vulnerability of classification systems with high-dimensional categorical inputs." @default.
- W4311640020 created "2022-12-27" @default.
- W4311640020 creator A5000024258 @default.
- W4311640020 creator A5000755750 @default.
- W4311640020 creator A5004680428 @default.
- W4311640020 creator A5007273011 @default.
- W4311640020 creator A5069664352 @default.
- W4311640020 date "2022-12-13" @default.
- W4311640020 modified "2023-09-30" @default.
- W4311640020 title "Towards Efficient and Domain-Agnostic Evasion Attack with High-dimensional Categorical Inputs" @default.
- W4311640020 doi "https://doi.org/10.48550/arxiv.2212.06836" @default.
- W4311640020 hasPublicationYear "2022" @default.
- W4311640020 type Work @default.
- W4311640020 citedByCount "0" @default.
- W4311640020 crossrefType "posted-content" @default.
- W4311640020 hasAuthorship W4311640020A5000024258 @default.
- W4311640020 hasAuthorship W4311640020A5000755750 @default.
- W4311640020 hasAuthorship W4311640020A5004680428 @default.
- W4311640020 hasAuthorship W4311640020A5007273011 @default.
- W4311640020 hasAuthorship W4311640020A5069664352 @default.
- W4311640020 hasBestOaLocation W43116400201 @default.
- W4311640020 hasConcept C119857082 @default.
- W4311640020 hasConcept C134306372 @default.
- W4311640020 hasConcept C154945302 @default.
- W4311640020 hasConcept C33923547 @default.
- W4311640020 hasConcept C36503486 @default.
- W4311640020 hasConcept C41008148 @default.
- W4311640020 hasConcept C50817715 @default.
- W4311640020 hasConcept C5274069 @default.
- W4311640020 hasConceptScore W4311640020C119857082 @default.
- W4311640020 hasConceptScore W4311640020C134306372 @default.
- W4311640020 hasConceptScore W4311640020C154945302 @default.
- W4311640020 hasConceptScore W4311640020C33923547 @default.
- W4311640020 hasConceptScore W4311640020C36503486 @default.
- W4311640020 hasConceptScore W4311640020C41008148 @default.
- W4311640020 hasConceptScore W4311640020C50817715 @default.
- W4311640020 hasConceptScore W4311640020C5274069 @default.
- W4311640020 hasLocation W43116400201 @default.
- W4311640020 hasOpenAccess W4311640020 @default.
- W4311640020 hasPrimaryLocation W43116400201 @default.
- W4311640020 hasRelatedWork W2961085424 @default.
- W4311640020 hasRelatedWork W3046775127 @default.
- W4311640020 hasRelatedWork W3209574120 @default.
- W4311640020 hasRelatedWork W4205958290 @default.
- W4311640020 hasRelatedWork W4220812971 @default.
- W4311640020 hasRelatedWork W4286629047 @default.
- W4311640020 hasRelatedWork W4306321456 @default.
- W4311640020 hasRelatedWork W4306674287 @default.
- W4311640020 hasRelatedWork W86096423 @default.
- W4311640020 hasRelatedWork W4224009465 @default.
- W4311640020 isParatext "false" @default.
- W4311640020 isRetracted "false" @default.
- W4311640020 workType "article" @default.