Matches in SemOpenAlex for { <https://semopenalex.org/work/W2941441145> ?p ?o ?g. }
- W2941441145 abstract "The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency." @default.
- W2941441145 created "2019-05-03" @default.
- W2941441145 creator A5004096199 @default.
- W2941441145 creator A5023786468 @default.
- W2941441145 creator A5029105520 @default.
- W2941441145 date "2019-04-25" @default.
- W2941441145 modified "2023-09-27" @default.
- W2941441145 title "Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields." @default.
- W2941441145 cites W1766888123 @default.
- W2941441145 cites W1860887978 @default.
- W2941441145 cites W1972790352 @default.
- W2941441145 cites W1973976434 @default.
- W2941441145 cites W1991605728 @default.
- W2941441145 cites W2022735534 @default.
- W2941441145 cites W2072072671 @default.
- W2941441145 cites W2101926813 @default.
- W2941441145 cites W2103504761 @default.
- W2941441145 cites W2109200236 @default.
- W2941441145 cites W2109812093 @default.
- W2941441145 cites W2112328181 @default.
- W2941441145 cites W2112796928 @default.
- W2941441145 cites W2115441154 @default.
- W2941441145 cites W2151103935 @default.
- W2941441145 cites W2167383966 @default.
- W2941441145 cites W2194775991 @default.
- W2941441145 cites W2212384750 @default.
- W2941441145 cites W2340897893 @default.
- W2941441145 cites W2395611524 @default.
- W2941441145 cites W2412782625 @default.
- W2941441145 cites W2414711238 @default.
- W2941441145 cites W2504108613 @default.
- W2941441145 cites W252252322 @default.
- W2941441145 cites W2556967412 @default.
- W2941441145 cites W2560023338 @default.
- W2941441145 cites W2563709626 @default.
- W2941441145 cites W2565639579 @default.
- W2941441145 cites W2592939477 @default.
- W2941441145 cites W2601564443 @default.
- W2941441145 cites W2798791840 @default.
- W2941441145 cites W2810075754 @default.
- W2941441145 cites W2949194345 @default.
- W2941441145 cites W2951655307 @default.
- W2941441145 cites W2952632681 @default.
- W2941441145 cites W2962750597 @default.
- W2941441145 cites W2962850830 @default.
- W2941441145 cites W2963073398 @default.
- W2941441145 cites W2963107874 @default.
- W2941441145 cites W2963323244 @default.
- W2941441145 cites W2963374479 @default.
- W2941441145 cites W2963840672 @default.
- W2941441145 cites W2964288706 @default.
- W2941441145 cites W3138516063 @default.
- W2941441145 cites W603908379 @default.
- W2941441145 hasPublicationYear "2019" @default.
- W2941441145 type Work @default.
- W2941441145 sameAs 2941441145 @default.
- W2941441145 citedByCount "16" @default.
- W2941441145 countsByYear W29414411452019 @default.
- W2941441145 countsByYear W29414411452020 @default.
- W2941441145 countsByYear W29414411452021 @default.
- W2941441145 crossrefType "posted-content" @default.
- W2941441145 hasAuthorship W2941441145A5004096199 @default.
- W2941441145 hasAuthorship W2941441145A5023786468 @default.
- W2941441145 hasAuthorship W2941441145A5029105520 @default.
- W2941441145 hasConcept C11413529 @default.
- W2941441145 hasConcept C121332964 @default.
- W2941441145 hasConcept C138885662 @default.
- W2941441145 hasConcept C153180895 @default.
- W2941441145 hasConcept C154945302 @default.
- W2941441145 hasConcept C163716315 @default.
- W2941441145 hasConcept C19071747 @default.
- W2941441145 hasConcept C2776214188 @default.
- W2941441145 hasConcept C2778334786 @default.
- W2941441145 hasConcept C2779808786 @default.
- W2941441145 hasConcept C41008148 @default.
- W2941441145 hasConcept C41895202 @default.
- W2941441145 hasConcept C44870925 @default.
- W2941441145 hasConcept C62520636 @default.
- W2941441145 hasConcept C89600930 @default.
- W2941441145 hasConceptScore W2941441145C11413529 @default.
- W2941441145 hasConceptScore W2941441145C121332964 @default.
- W2941441145 hasConceptScore W2941441145C138885662 @default.
- W2941441145 hasConceptScore W2941441145C153180895 @default.
- W2941441145 hasConceptScore W2941441145C154945302 @default.
- W2941441145 hasConceptScore W2941441145C163716315 @default.
- W2941441145 hasConceptScore W2941441145C19071747 @default.
- W2941441145 hasConceptScore W2941441145C2776214188 @default.
- W2941441145 hasConceptScore W2941441145C2778334786 @default.
- W2941441145 hasConceptScore W2941441145C2779808786 @default.
- W2941441145 hasConceptScore W2941441145C41008148 @default.
- W2941441145 hasConceptScore W2941441145C41895202 @default.
- W2941441145 hasConceptScore W2941441145C44870925 @default.
- W2941441145 hasConceptScore W2941441145C62520636 @default.
- W2941441145 hasConceptScore W2941441145C89600930 @default.
- W2941441145 hasLocation W29414411451 @default.
- W2941441145 hasOpenAccess W2941441145 @default.
- W2941441145 hasPrimaryLocation W29414411451 @default.
- W2941441145 hasRelatedWork W1861492603 @default.
- W2941441145 hasRelatedWork W1973976434 @default.
- W2941441145 hasRelatedWork W2022735534 @default.