Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384663241> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4384663241 endingPage "9446" @default.
- W4384663241 startingPage "9432" @default.
- W4384663241 abstract "The many successes of deep neural networks (DNNs) over the past decade have largely been driven by computational scale rather than insights from biological intelligence. Here, we explore if these trends have also carried concomitant improvements in explaining the visual strategies humans rely on for object recognition. We do this by comparing two related but distinct properties of visual strategies in humans and DNNs: where they believe important visual features are in images and how they use those features to categorize objects. Across 84 different DNNs trained on ImageNet and three independent datasets measuring the where and the how of human visual strategies for object recognition on those images, we find a systematic trade-off between DNN categorization accuracy and alignment with human visual strategies for object recognition. State-of-the-art DNNs are progressively becoming less aligned with humans as their accuracy improves. We rectify this growing issue with our neural harmonizer: a general-purpose training routine that both aligns DNN and human visual strategies and improves categorization accuracy. Our work represents the first demonstration that the scaling laws [1-3] that are guiding the design of DNNs today have also produced worse models of human vision. We release our code and data at https://serre-lab.github.io/Harmonization to help the field build more human-like DNNs." @default.
- W4384663241 created "2023-07-20" @default.
- W4384663241 creator A5018626055 @default.
- W4384663241 creator A5036483775 @default.
- W4384663241 creator A5038564554 @default.
- W4384663241 creator A5051681063 @default.
- W4384663241 date "2022-12-01" @default.
- W4384663241 modified "2023-10-13" @default.
- W4384663241 title "Harmonizing the object recognition strategies of deep neural networks with humans." @default.
- W4384663241 cites W1994658111 @default.
- W4384663241 cites W1997069228 @default.
- W4384663241 cites W2058616551 @default.
- W4384663241 cites W2103410189 @default.
- W4384663241 cites W2122896223 @default.
- W4384663241 cites W2129673518 @default.
- W4384663241 cites W2132172482 @default.
- W4384663241 cites W2162950292 @default.
- W4384663241 cites W2166198088 @default.
- W4384663241 cites W2280426979 @default.
- W4384663241 cites W2623160143 @default.
- W4384663241 cites W2795373741 @default.
- W4384663241 cites W2803790834 @default.
- W4384663241 cites W2898200825 @default.
- W4384663241 cites W2928165649 @default.
- W4384663241 cites W2966900272 @default.
- W4384663241 cites W2979357328 @default.
- W4384663241 cites W3086360455 @default.
- W4384663241 cites W3099595508 @default.
- W4384663241 cites W3108490213 @default.
- W4384663241 cites W3146449001 @default.
- W4384663241 cites W3193620486 @default.
- W4384663241 cites W4225585129 @default.
- W4384663241 cites W4226318754 @default.
- W4384663241 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37465369" @default.
- W4384663241 hasPublicationYear "2022" @default.
- W4384663241 type Work @default.
- W4384663241 citedByCount "1" @default.
- W4384663241 crossrefType "journal-article" @default.
- W4384663241 hasAuthorship W4384663241A5018626055 @default.
- W4384663241 hasAuthorship W4384663241A5036483775 @default.
- W4384663241 hasAuthorship W4384663241A5038564554 @default.
- W4384663241 hasAuthorship W4384663241A5051681063 @default.
- W4384663241 hasConcept C119857082 @default.
- W4384663241 hasConcept C121332964 @default.
- W4384663241 hasConcept C154945302 @default.
- W4384663241 hasConcept C24890656 @default.
- W4384663241 hasConcept C2779962950 @default.
- W4384663241 hasConcept C2781238097 @default.
- W4384663241 hasConcept C2984842247 @default.
- W4384663241 hasConcept C41008148 @default.
- W4384663241 hasConcept C50644808 @default.
- W4384663241 hasConcept C64876066 @default.
- W4384663241 hasConcept C94124525 @default.
- W4384663241 hasConceptScore W4384663241C119857082 @default.
- W4384663241 hasConceptScore W4384663241C121332964 @default.
- W4384663241 hasConceptScore W4384663241C154945302 @default.
- W4384663241 hasConceptScore W4384663241C24890656 @default.
- W4384663241 hasConceptScore W4384663241C2779962950 @default.
- W4384663241 hasConceptScore W4384663241C2781238097 @default.
- W4384663241 hasConceptScore W4384663241C2984842247 @default.
- W4384663241 hasConceptScore W4384663241C41008148 @default.
- W4384663241 hasConceptScore W4384663241C50644808 @default.
- W4384663241 hasConceptScore W4384663241C64876066 @default.
- W4384663241 hasConceptScore W4384663241C94124525 @default.
- W4384663241 hasLocation W43846632411 @default.
- W4384663241 hasOpenAccess W4384663241 @default.
- W4384663241 hasPrimaryLocation W43846632411 @default.
- W4384663241 hasRelatedWork W1994959803 @default.
- W4384663241 hasRelatedWork W2109586375 @default.
- W4384663241 hasRelatedWork W2115592544 @default.
- W4384663241 hasRelatedWork W2141790984 @default.
- W4384663241 hasRelatedWork W2142935099 @default.
- W4384663241 hasRelatedWork W2652818544 @default.
- W4384663241 hasRelatedWork W4308756403 @default.
- W4384663241 hasRelatedWork W4365135359 @default.
- W4384663241 hasRelatedWork W4380086463 @default.
- W4384663241 hasRelatedWork W1629725936 @default.
- W4384663241 hasVolume "35" @default.
- W4384663241 isParatext "false" @default.
- W4384663241 isRetracted "false" @default.
- W4384663241 workType "article" @default.