Matches in SemOpenAlex for { <https://semopenalex.org/work/W3007511799> ?p ?o ?g. }
- W3007511799 abstract "Intensive algorithmic efforts have been made to enable the rapid improvements of certificated robustness for complex ML models recently. However, current robustness certification methods are only able to certify under a limited perturbation radius. Given that existing pure data-driven statistical approaches have reached a bottleneck, in this paper, we propose to integrate statistical ML models with knowledge (expressed as logical rules) as a reasoning component using Markov logic networks (MLN, so as to further improve the overall certified robustness. This opens new research questions about certifying the robustness of such a paradigm, especially the reasoning component (e.g., MLN). As the first step towards understanding these questions, we first prove that the computational complexity of certifying the robustness of MLN is #P-hard. Guided by this hardness result, we then derive the first certified robustness bound for MLN by carefully analyzing different model regimes. Finally, we conduct extensive experiments on five datasets including both high-dimensional images and natural language texts, and we show that the certified robustness with knowledge-based logical reasoning indeed significantly outperforms that of the state-of-the-arts." @default.
- W3007511799 created "2020-03-06" @default.
- W3007511799 creator A5022706581 @default.
- W3007511799 creator A5025073517 @default.
- W3007511799 creator A5034995105 @default.
- W3007511799 creator A5036100753 @default.
- W3007511799 creator A5041191241 @default.
- W3007511799 creator A5041232755 @default.
- W3007511799 creator A5044586758 @default.
- W3007511799 creator A5049225160 @default.
- W3007511799 creator A5052515058 @default.
- W3007511799 date "2020-02-28" @default.
- W3007511799 modified "2023-10-16" @default.
- W3007511799 title "Improving Certified Robustness via Statistical Learning with Logical Reasoning" @default.
- W3007511799 cites W1482212161 @default.
- W3007511799 cites W1535439311 @default.
- W3007511799 cites W1599188306 @default.
- W3007511799 cites W1977970897 @default.
- W3007511799 cites W1993924277 @default.
- W3007511799 cites W2052569640 @default.
- W3007511799 cites W2052963087 @default.
- W3007511799 cites W2096669529 @default.
- W3007511799 cites W2098679902 @default.
- W3007511799 cites W2108598243 @default.
- W3007511799 cites W2109394932 @default.
- W3007511799 cites W2120340025 @default.
- W3007511799 cites W2126267628 @default.
- W3007511799 cites W2133291805 @default.
- W3007511799 cites W2166741250 @default.
- W3007511799 cites W2408141691 @default.
- W3007511799 cites W2608353553 @default.
- W3007511799 cites W2746600820 @default.
- W3007511799 cites W2768915615 @default.
- W3007511799 cites W2783555701 @default.
- W3007511799 cites W2798302089 @default.
- W3007511799 cites W2807040120 @default.
- W3007511799 cites W2895595267 @default.
- W3007511799 cites W2896457183 @default.
- W3007511799 cites W2897355816 @default.
- W3007511799 cites W2898963688 @default.
- W3007511799 cites W2911634294 @default.
- W3007511799 cites W2923722540 @default.
- W3007511799 cites W2949362401 @default.
- W3007511799 cites W2950864148 @default.
- W3007511799 cites W2963070423 @default.
- W3007511799 cites W2963207607 @default.
- W3007511799 cites W2963424284 @default.
- W3007511799 cites W2963440492 @default.
- W3007511799 cites W2963496101 @default.
- W3007511799 cites W2963744840 @default.
- W3007511799 cites W2963857521 @default.
- W3007511799 cites W2964253222 @default.
- W3007511799 cites W2965677914 @default.
- W3007511799 cites W2970680991 @default.
- W3007511799 cites W2971109239 @default.
- W3007511799 cites W3026969282 @default.
- W3007511799 cites W3037755290 @default.
- W3007511799 cites W3083878034 @default.
- W3007511799 cites W3107942585 @default.
- W3007511799 cites W3133236490 @default.
- W3007511799 cites W3133465759 @default.
- W3007511799 cites W3163244177 @default.
- W3007511799 cites W64813323 @default.
- W3007511799 doi "https://doi.org/10.48550/arxiv.2003.00120" @default.
- W3007511799 hasPublicationYear "2020" @default.
- W3007511799 type Work @default.
- W3007511799 sameAs 3007511799 @default.
- W3007511799 citedByCount "3" @default.
- W3007511799 countsByYear W30075117992020 @default.
- W3007511799 countsByYear W30075117992021 @default.
- W3007511799 crossrefType "posted-content" @default.
- W3007511799 hasAuthorship W3007511799A5022706581 @default.
- W3007511799 hasAuthorship W3007511799A5025073517 @default.
- W3007511799 hasAuthorship W3007511799A5034995105 @default.
- W3007511799 hasAuthorship W3007511799A5036100753 @default.
- W3007511799 hasAuthorship W3007511799A5041191241 @default.
- W3007511799 hasAuthorship W3007511799A5041232755 @default.
- W3007511799 hasAuthorship W3007511799A5044586758 @default.
- W3007511799 hasAuthorship W3007511799A5049225160 @default.
- W3007511799 hasAuthorship W3007511799A5052515058 @default.
- W3007511799 hasBestOaLocation W30075117991 @default.
- W3007511799 hasConcept C104317684 @default.
- W3007511799 hasConcept C119857082 @default.
- W3007511799 hasConcept C149635348 @default.
- W3007511799 hasConcept C154945302 @default.
- W3007511799 hasConcept C17744445 @default.
- W3007511799 hasConcept C185592680 @default.
- W3007511799 hasConcept C199539241 @default.
- W3007511799 hasConcept C2780513914 @default.
- W3007511799 hasConcept C41008148 @default.
- W3007511799 hasConcept C46304622 @default.
- W3007511799 hasConcept C55493867 @default.
- W3007511799 hasConcept C63479239 @default.
- W3007511799 hasConcept C80444323 @default.
- W3007511799 hasConceptScore W3007511799C104317684 @default.
- W3007511799 hasConceptScore W3007511799C119857082 @default.
- W3007511799 hasConceptScore W3007511799C149635348 @default.
- W3007511799 hasConceptScore W3007511799C154945302 @default.
- W3007511799 hasConceptScore W3007511799C17744445 @default.
- W3007511799 hasConceptScore W3007511799C185592680 @default.