FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3106646540> ?p ?o ?g. }

• W3106646540 abstract "Self-supervised visual representation learning has seen huge progress recently, but no large scale evaluation has compared the many models now available. We evaluate the transfer performance of 13 top self-supervised models on 40 downstream tasks, including many-shot and few-shot recognition, object detection, and dense prediction. We compare their performance to a supervised baseline and show that on most tasks the best self-supervised models outperform supervision, confirming the recently observed trend in the literature. We find ImageNet Top-1 accuracy to be highly correlated with transfer to many-shot recognition, but increasingly less so for few-shot, object detection and dense prediction. No single self-supervised method dominates overall, suggesting that universal pre-training is still unsolved. Our analysis of features suggests that top self-supervised learners fail to preserve colour information as well as supervised alternatives, but tend to induce better classifier calibration, and less attentive overfitting than supervised learners." @default.
• W3106646540 created "2020-12-07" @default.
• W3106646540 creator A5039741908 @default.
• W3106646540 creator A5082476745 @default.
• W3106646540 creator A5087823932 @default.
• W3106646540 date "2020-11-26" @default.
• W3106646540 modified "2023-09-27" @default.
• W3106646540 title "How Well Do Self-Supervised Models Transfer?" @default.
• W3106646540 cites W125693051 @default.
• W3106646540 cites W12634471 @default.
• W3106646540 cites W1846799578 @default.
• W3106646540 cites W1849277567 @default.
• W3106646540 cites W1977295328 @default.
• W3106646540 cites W1993309459 @default.
• W3106646540 cites W2017814585 @default.
• W3106646540 cites W2031489346 @default.
• W3106646540 cites W2047643928 @default.
• W3106646540 cites W2051434435 @default.
• W3106646540 cites W2108598243 @default.
• W3106646540 cites W2152790380 @default.
• W3106646540 cites W2155904486 @default.
• W3106646540 cites W2194775991 @default.
• W3106646540 cites W2308529009 @default.
• W3106646540 cites W2321533354 @default.
• W3106646540 cites W2326925005 @default.
• W3106646540 cites W2473156356 @default.
• W3106646540 cites W2507296351 @default.
• W3106646540 cites W2533598788 @default.
• W3106646540 cites W2560023338 @default.
• W3106646540 cites W2561238782 @default.
• W3106646540 cites W2565639579 @default.
• W3106646540 cites W2601450892 @default.
• W3106646540 cites W2613718673 @default.
• W3106646540 cites W2783991250 @default.
• W3106646540 cites W2798991696 @default.
• W3106646540 cites W2804935296 @default.
• W3106646540 cites W2883725317 @default.
• W3106646540 cites W2884822772 @default.
• W3106646540 cites W2913939497 @default.
• W3106646540 cites W2932083555 @default.
• W3106646540 cites W2950794831 @default.
• W3106646540 cites W2962785568 @default.
• W3106646540 cites W2964013315 @default.
• W3106646540 cites W2964194231 @default.
• W3106646540 cites W2964212410 @default.
• W3106646540 cites W2970854840 @default.
• W3106646540 cites W2970971581 @default.
• W3106646540 cites W2990873191 @default.
• W3106646540 cites W2991391304 @default.
• W3106646540 cites W2995181141 @default.
• W3106646540 cites W3008526508 @default.
• W3106646540 cites W3009561768 @default.
• W3106646540 cites W3022061250 @default.
• W3106646540 cites W3023371261 @default.
• W3106646540 cites W3026414903 @default.
• W3106646540 cites W3034576826 @default.
• W3106646540 cites W3034781633 @default.
• W3106646540 cites W3034978746 @default.
• W3106646540 cites W3035058308 @default.
• W3106646540 cites W3035447285 @default.
• W3106646540 cites W3035524453 @default.
• W3106646540 cites W3089824566 @default.
• W3106646540 cites W3095121901 @default.
• W3106646540 cites W3100859887 @default.
• W3106646540 cites W3101156210 @default.
• W3106646540 cites W3101821705 @default.
• W3106646540 cites W3102785203 @default.
• W3106646540 cites W3106428938 @default.
• W3106646540 cites W3108655343 @default.
• W3106646540 cites W3110608229 @default.
• W3106646540 cites W3118608800 @default.
• W3106646540 cites W3122325173 @default.
• W3106646540 cites W3123072794 @default.
• W3106646540 cites W337610345 @default.
• W3106646540 cites W780950768 @default.
• W3106646540 hasPublicationYear "2020" @default.
• W3106646540 type Work @default.
• W3106646540 sameAs 3106646540 @default.
• W3106646540 citedByCount "2" @default.
• W3106646540 countsByYear W31066465402021 @default.
• W3106646540 crossrefType "posted-content" @default.
• W3106646540 hasAuthorship W3106646540A5039741908 @default.
• W3106646540 hasAuthorship W3106646540A5082476745 @default.
• W3106646540 hasAuthorship W3106646540A5087823932 @default.
• W3106646540 hasConcept C119857082 @default.
• W3106646540 hasConcept C136389625 @default.
• W3106646540 hasConcept C150899416 @default.
• W3106646540 hasConcept C153180895 @default.
• W3106646540 hasConcept C154945302 @default.
• W3106646540 hasConcept C22019652 @default.
• W3106646540 hasConcept C2781238097 @default.
• W3106646540 hasConcept C41008148 @default.
• W3106646540 hasConcept C50644808 @default.
• W3106646540 hasConcept C95623464 @default.
• W3106646540 hasConceptScore W3106646540C119857082 @default.
• W3106646540 hasConceptScore W3106646540C136389625 @default.
• W3106646540 hasConceptScore W3106646540C150899416 @default.
• W3106646540 hasConceptScore W3106646540C153180895 @default.
• W3106646540 hasConceptScore W3106646540C154945302 @default.
• W3106646540 hasConceptScore W3106646540C22019652 @default.

• next