FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037693414> ?p ?o ?g. }

• W3037693414 abstract "Given a collection of images, humans are able to discover landmarks of the depicted objects by modeling the shared geometric structure across instances. This idea of geometric equivariance has been widely used for unsupervised discovery of object landmark representations. In this paper, we develop a simple and effective approach based on contrastive learning of invariant representations. We show that when a deep network is trained to be invariant to geometric and photometric transformations, representations from its intermediate layers are highly predictive of object landmarks. Furthermore, by stacking representations across layers in a hypercolumn their effectiveness can be improved. Our approach is motivated by the phenomenon of the gradual emergence of invariance in the representation hierarchy of a deep network. We also present a unified view of existing equivariant and invariant representation learning approaches through the lens of contrastive learning, shedding light on the nature of invariances learned. Experiments on standard benchmarks for landmark discovery, as well as a challenging one we propose, show that the proposed approach surpasses prior state-of-the-art." @default.
• W3037693414 created "2020-07-02" @default.
• W3037693414 creator A5046487166 @default.
• W3037693414 creator A5052551454 @default.
• W3037693414 creator A5067219074 @default.
• W3037693414 date "2020-06-26" @default.
• W3037693414 modified "2023-09-27" @default.
• W3037693414 title "Unsupervised Discovery of Object Landmarks via Contrastive Learning" @default.
• W3037693414 cites W1522301498 @default.
• W3037693414 cites W1724369340 @default.
• W3037693414 cites W1797268635 @default.
• W3037693414 cites W1849277567 @default.
• W3037693414 cites W1909952827 @default.
• W3037693414 cites W1928906481 @default.
• W3037693414 cites W1948751323 @default.
• W3037693414 cites W1994488211 @default.
• W3037693414 cites W2012885984 @default.
• W3037693414 cites W2131846894 @default.
• W3037693414 cites W2133774033 @default.
• W3037693414 cites W2136026194 @default.
• W3037693414 cites W2138621090 @default.
• W3037693414 cites W2148349024 @default.
• W3037693414 cites W2151103935 @default.
• W3037693414 cites W2152790380 @default.
• W3037693414 cites W2157558673 @default.
• W3037693414 cites W2161969291 @default.
• W3037693414 cites W2190008860 @default.
• W3037693414 cites W2321533354 @default.
• W3037693414 cites W2611103765 @default.
• W3037693414 cites W2737725206 @default.
• W3037693414 cites W2770121394 @default.
• W3037693414 cites W2770302340 @default.
• W3037693414 cites W2798991696 @default.
• W3037693414 cites W2842511635 @default.
• W3037693414 cites W2887997593 @default.
• W3037693414 cites W2902227449 @default.
• W3037693414 cites W2922461432 @default.
• W3037693414 cites W2935908327 @default.
• W3037693414 cites W2943527153 @default.
• W3037693414 cites W2944828972 @default.
• W3037693414 cites W2948012107 @default.
• W3037693414 cites W2949074271 @default.
• W3037693414 cites W2949517790 @default.
• W3037693414 cites W2949650786 @default.
• W3037693414 cites W2949678110 @default.
• W3037693414 cites W2949886837 @default.
• W3037693414 cites W2950328304 @default.
• W3037693414 cites W2951019013 @default.
• W3037693414 cites W2952054889 @default.
• W3037693414 cites W2962742544 @default.
• W3037693414 cites W2963203586 @default.
• W3037693414 cites W2963419579 @default.
• W3037693414 cites W2987283559 @default.
• W3037693414 cites W2995489995 @default.
• W3037693414 cites W3005680577 @default.
• W3037693414 cites W3009561768 @default.
• W3037693414 cites W3113653931 @default.
• W3037693414 hasPublicationYear "2020" @default.
• W3037693414 type Work @default.
• W3037693414 sameAs 3037693414 @default.
• W3037693414 citedByCount "3" @default.
• W3037693414 countsByYear W30376934142021 @default.
• W3037693414 crossrefType "posted-content" @default.
• W3037693414 hasAuthorship W3037693414A5046487166 @default.
• W3037693414 hasAuthorship W3037693414A5052551454 @default.
• W3037693414 hasAuthorship W3037693414A5067219074 @default.
• W3037693414 hasConcept C108583219 @default.
• W3037693414 hasConcept C153180895 @default.
• W3037693414 hasConcept C154945302 @default.
• W3037693414 hasConcept C171036898 @default.
• W3037693414 hasConcept C17744445 @default.
• W3037693414 hasConcept C190470478 @default.
• W3037693414 hasConcept C199539241 @default.
• W3037693414 hasConcept C202444582 @default.
• W3037693414 hasConcept C2776359362 @default.
• W3037693414 hasConcept C2780297707 @default.
• W3037693414 hasConcept C2781238097 @default.
• W3037693414 hasConcept C33923547 @default.
• W3037693414 hasConcept C37914503 @default.
• W3037693414 hasConcept C41008148 @default.
• W3037693414 hasConcept C59404180 @default.
• W3037693414 hasConcept C8038995 @default.
• W3037693414 hasConcept C94625758 @default.
• W3037693414 hasConceptScore W3037693414C108583219 @default.
• W3037693414 hasConceptScore W3037693414C153180895 @default.
• W3037693414 hasConceptScore W3037693414C154945302 @default.
• W3037693414 hasConceptScore W3037693414C171036898 @default.
• W3037693414 hasConceptScore W3037693414C17744445 @default.
• W3037693414 hasConceptScore W3037693414C190470478 @default.
• W3037693414 hasConceptScore W3037693414C199539241 @default.
• W3037693414 hasConceptScore W3037693414C202444582 @default.
• W3037693414 hasConceptScore W3037693414C2776359362 @default.
• W3037693414 hasConceptScore W3037693414C2780297707 @default.
• W3037693414 hasConceptScore W3037693414C2781238097 @default.
• W3037693414 hasConceptScore W3037693414C33923547 @default.
• W3037693414 hasConceptScore W3037693414C37914503 @default.
• W3037693414 hasConceptScore W3037693414C41008148 @default.
• W3037693414 hasConceptScore W3037693414C59404180 @default.
• W3037693414 hasConceptScore W3037693414C8038995 @default.
• W3037693414 hasConceptScore W3037693414C94625758 @default.

• next