Matches in SemOpenAlex for { <https://semopenalex.org/work/W2951174316> ?p ?o ?g. }
- W2951174316 abstract "We study the problem of grounding distributional representations of texts on the visual domain, namely visual-semantic embeddings (VSE for short). Begin with an insightful adversarial attack on VSE embeddings, we show the limitation of current frameworks and image-text datasets (e.g., MS-COCO) both quantitatively and qualitatively. The large gap between the number of possible constitutions of real-world semantics and the size of parallel data, to a large extent, restricts the model to establish the link between textual semantics and visual concepts. We alleviate this problem by augmenting the MS-COCO image captioning datasets with textual contrastive adversarial samples. These samples are synthesized using linguistic rules and the WordNet knowledge base. The construction procedure is both syntax- and semantics-aware. The samples enforce the model to ground learned embeddings to concrete concepts within the image. This simple but powerful technique brings a noticeable improvement over the baselines on a diverse set of downstream tasks, in addition to defending known-type adversarial attacks. We release the codes at this https URL." @default.
- W2951174316 created "2019-06-27" @default.
- W2951174316 creator A5004183775 @default.
- W2951174316 creator A5033435691 @default.
- W2951174316 creator A5041246275 @default.
- W2951174316 creator A5074655314 @default.
- W2951174316 creator A5091670413 @default.
- W2951174316 date "2018-06-27" @default.
- W2951174316 modified "2023-09-27" @default.
- W2951174316 title "Learning Visually-Grounded Semantics from Contrastive Adversarial Samples" @default.
- W2951174316 cites W1523385540 @default.
- W2951174316 cites W1527575280 @default.
- W2951174316 cites W1861492603 @default.
- W2951174316 cites W1905882502 @default.
- W2951174316 cites W1932198206 @default.
- W2951174316 cites W2009077327 @default.
- W2951174316 cites W2025341678 @default.
- W2951174316 cites W2081580037 @default.
- W2951174316 cites W2118090838 @default.
- W2951174316 cites W2123024445 @default.
- W2951174316 cites W2149557440 @default.
- W2951174316 cites W2169783907 @default.
- W2951174316 cites W2184188583 @default.
- W2951174316 cites W2185175083 @default.
- W2951174316 cites W2194775991 @default.
- W2951174316 cites W2250539671 @default.
- W2951174316 cites W2251329024 @default.
- W2951174316 cites W2411480514 @default.
- W2951174316 cites W2474574787 @default.
- W2951174316 cites W2521709538 @default.
- W2951174316 cites W2546696630 @default.
- W2951174316 cites W2552579943 @default.
- W2951174316 cites W2616841723 @default.
- W2951174316 cites W2738015883 @default.
- W2951174316 cites W2756386045 @default.
- W2951174316 cites W2774267535 @default.
- W2951174316 cites W2778940641 @default.
- W2951174316 cites W2950577311 @default.
- W2951174316 cites W2950774971 @default.
- W2951174316 cites W2951597448 @default.
- W2951174316 cites W2952191002 @default.
- W2951174316 cites W2952246170 @default.
- W2951174316 cites W2962964385 @default.
- W2951174316 cites W2963389687 @default.
- W2951174316 cites W2964091467 @default.
- W2951174316 cites W2964121744 @default.
- W2951174316 hasPublicationYear "2018" @default.
- W2951174316 type Work @default.
- W2951174316 sameAs 2951174316 @default.
- W2951174316 citedByCount "3" @default.
- W2951174316 countsByYear W29511743162018 @default.
- W2951174316 countsByYear W29511743162020 @default.
- W2951174316 crossrefType "posted-content" @default.
- W2951174316 hasAuthorship W2951174316A5004183775 @default.
- W2951174316 hasAuthorship W2951174316A5033435691 @default.
- W2951174316 hasAuthorship W2951174316A5041246275 @default.
- W2951174316 hasAuthorship W2951174316A5074655314 @default.
- W2951174316 hasAuthorship W2951174316A5091670413 @default.
- W2951174316 hasConcept C115961682 @default.
- W2951174316 hasConcept C134306372 @default.
- W2951174316 hasConcept C154945302 @default.
- W2951174316 hasConcept C157657479 @default.
- W2951174316 hasConcept C157659113 @default.
- W2951174316 hasConcept C177264268 @default.
- W2951174316 hasConcept C184337299 @default.
- W2951174316 hasConcept C199360897 @default.
- W2951174316 hasConcept C204321447 @default.
- W2951174316 hasConcept C33923547 @default.
- W2951174316 hasConcept C36503486 @default.
- W2951174316 hasConcept C37736160 @default.
- W2951174316 hasConcept C41008148 @default.
- W2951174316 hasConcept C60048249 @default.
- W2951174316 hasConceptScore W2951174316C115961682 @default.
- W2951174316 hasConceptScore W2951174316C134306372 @default.
- W2951174316 hasConceptScore W2951174316C154945302 @default.
- W2951174316 hasConceptScore W2951174316C157657479 @default.
- W2951174316 hasConceptScore W2951174316C157659113 @default.
- W2951174316 hasConceptScore W2951174316C177264268 @default.
- W2951174316 hasConceptScore W2951174316C184337299 @default.
- W2951174316 hasConceptScore W2951174316C199360897 @default.
- W2951174316 hasConceptScore W2951174316C204321447 @default.
- W2951174316 hasConceptScore W2951174316C33923547 @default.
- W2951174316 hasConceptScore W2951174316C36503486 @default.
- W2951174316 hasConceptScore W2951174316C37736160 @default.
- W2951174316 hasConceptScore W2951174316C41008148 @default.
- W2951174316 hasConceptScore W2951174316C60048249 @default.
- W2951174316 hasLocation W29511743161 @default.
- W2951174316 hasOpenAccess W2951174316 @default.
- W2951174316 hasPrimaryLocation W29511743161 @default.
- W2951174316 hasRelatedWork W2096831027 @default.
- W2951174316 hasRelatedWork W2110182048 @default.
- W2951174316 hasRelatedWork W2517456239 @default.
- W2951174316 hasRelatedWork W2567035470 @default.
- W2951174316 hasRelatedWork W2594230509 @default.
- W2951174316 hasRelatedWork W2789718859 @default.
- W2951174316 hasRelatedWork W2810689564 @default.
- W2951174316 hasRelatedWork W2895466885 @default.
- W2951174316 hasRelatedWork W2905190525 @default.
- W2951174316 hasRelatedWork W2913009085 @default.
- W2951174316 hasRelatedWork W2953591834 @default.