FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2996098493> ?p ?o ?g. }

• W2996098493 abstract "Analyzing the behavior of neural networks is one of the most pressing challenges in deep learning. Binarized Neural Networks are an important class of networks that allow equivalent representation in Boolean logic and can be analyzed formally with logic-based reasoning tools like SAT solvers. Such tools can be used to answer existential and probabilistic queries about the network, perform explanation generation, etc. However, the main bottleneck for all methods is their ability to reason about large BNNs efficiently. In this work, we analyze architectural design choices of BNNs and discuss how they affect the performance of logic-based reasoners. We propose changes to the BNN architecture and the training procedure to get a simpler network for SAT solvers without sacrificing accuracy on the primary task. Our experimental results demonstrate that our approach scales to larger deep neural networks compared to existing work for existential and probabilistic queries, leading to significant speed ups on all tested datasets." @default.
• W2996098493 created "2019-12-26" @default.
• W2996098493 creator A5003614499 @default.
• W2996098493 creator A5034897040 @default.
• W2996098493 creator A5041231710 @default.
• W2996098493 creator A5072902302 @default.
• W2996098493 date "2020-04-30" @default.
• W2996098493 modified "2023-10-16" @default.
• W2996098493 title "In Search for a SAT-friendly Binarized Neural Network Architecture" @default.
• W2996098493 cites W1537730723 @default.
• W2996098493 cites W1549166962 @default.
• W2996098493 cites W1605798759 @default.
• W2996098493 cites W1758913682 @default.
• W2996098493 cites W189719929 @default.
• W2996098493 cites W2019363670 @default.
• W2996098493 cites W2100609826 @default.
• W2996098493 cites W2115826669 @default.
• W2996098493 cites W2126111067 @default.
• W2996098493 cites W2186752618 @default.
• W2996098493 cites W2267635276 @default.
• W2996098493 cites W2493343568 @default.
• W2996098493 cites W2557283755 @default.
• W2996098493 cites W2594877703 @default.
• W2996098493 cites W2623381305 @default.
• W2996098493 cites W2741061345 @default.
• W2996098493 cites W2788645721 @default.
• W2996098493 cites W2803392236 @default.
• W2996098493 cites W2803850896 @default.
• W2996098493 cites W2807040120 @default.
• W2996098493 cites W2809729522 @default.
• W2996098493 cites W2890472662 @default.
• W2996098493 cites W2892354372 @default.
• W2996098493 cites W2895089194 @default.
• W2996098493 cites W2908314593 @default.
• W2996098493 cites W2909765152 @default.
• W2996098493 cites W2949117887 @default.
• W2996098493 cites W2953796333 @default.
• W2996098493 cites W2957311447 @default.
• W2996098493 cites W2963061092 @default.
• W2996098493 cites W2963207607 @default.
• W2996098493 cites W2963496101 @default.
• W2996098493 cites W2963673089 @default.
• W2996098493 cites W2963813662 @default.
• W2996098493 cites W2963983756 @default.
• W2996098493 cites W2964396926 @default.
• W2996098493 cites W2989358546 @default.
• W2996098493 cites W967948971 @default.
• W2996098493 hasPublicationYear "2020" @default.
• W2996098493 type Work @default.
• W2996098493 sameAs 2996098493 @default.
• W2996098493 citedByCount "6" @default.
• W2996098493 countsByYear W29960984932020 @default.
• W2996098493 countsByYear W29960984932021 @default.
• W2996098493 crossrefType "proceedings-article" @default.
• W2996098493 hasAuthorship W2996098493A5003614499 @default.
• W2996098493 hasAuthorship W2996098493A5034897040 @default.
• W2996098493 hasAuthorship W2996098493A5041231710 @default.
• W2996098493 hasAuthorship W2996098493A5072902302 @default.
• W2996098493 hasConcept C108583219 @default.
• W2996098493 hasConcept C119857082 @default.
• W2996098493 hasConcept C149635348 @default.
• W2996098493 hasConcept C154945302 @default.
• W2996098493 hasConcept C162324750 @default.
• W2996098493 hasConcept C17744445 @default.
• W2996098493 hasConcept C187736073 @default.
• W2996098493 hasConcept C199539241 @default.
• W2996098493 hasConcept C2776359362 @default.
• W2996098493 hasConcept C2777212361 @default.
• W2996098493 hasConcept C2780451532 @default.
• W2996098493 hasConcept C2780513914 @default.
• W2996098493 hasConcept C41008148 @default.
• W2996098493 hasConcept C49937458 @default.
• W2996098493 hasConcept C50644808 @default.
• W2996098493 hasConcept C80444323 @default.
• W2996098493 hasConcept C94625758 @default.
• W2996098493 hasConceptScore W2996098493C108583219 @default.
• W2996098493 hasConceptScore W2996098493C119857082 @default.
• W2996098493 hasConceptScore W2996098493C149635348 @default.
• W2996098493 hasConceptScore W2996098493C154945302 @default.
• W2996098493 hasConceptScore W2996098493C162324750 @default.
• W2996098493 hasConceptScore W2996098493C17744445 @default.
• W2996098493 hasConceptScore W2996098493C187736073 @default.
• W2996098493 hasConceptScore W2996098493C199539241 @default.
• W2996098493 hasConceptScore W2996098493C2776359362 @default.
• W2996098493 hasConceptScore W2996098493C2777212361 @default.
• W2996098493 hasConceptScore W2996098493C2780451532 @default.
• W2996098493 hasConceptScore W2996098493C2780513914 @default.
• W2996098493 hasConceptScore W2996098493C41008148 @default.
• W2996098493 hasConceptScore W2996098493C49937458 @default.
• W2996098493 hasConceptScore W2996098493C50644808 @default.
• W2996098493 hasConceptScore W2996098493C80444323 @default.
• W2996098493 hasConceptScore W2996098493C94625758 @default.
• W2996098493 hasLocation W29960984931 @default.
• W2996098493 hasOpenAccess W2996098493 @default.
• W2996098493 hasPrimaryLocation W29960984931 @default.
• W2996098493 hasRelatedWork W1516197167 @default.
• W2996098493 hasRelatedWork W1552673525 @default.
• W2996098493 hasRelatedWork W2267635276 @default.
• W2996098493 hasRelatedWork W2556644972 @default.
• W2996098493 hasRelatedWork W2559911651 @default.

• next