Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100776995> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W3100776995 abstract "Exploiting visual groundings for language understanding has recently been drawing much attention. In this work, we study visually grounded grammar induction and learn a constituency parser from both unlabeled text and its visual groundings. Existing work on this task (Shi et al., 2019) optimizes a parser via Reinforce and derives the learning signal only from the alignment of images and sentences. While their model is relatively accurate overall, its error distribution is very uneven, with low performance on certain constituents types (e.g., 26.2% recall on verb phrases, VPs) and high on others (e.g., 79.6% recall on noun phrases, NPs). This is not surprising as the learning signal is likely insufficient for deriving all aspects of phrase-structure syntax and gradient estimates are noisy. We show that using an extension of probabilistic context-free grammar model we can do fully-differentiable end-to-end visually grounded learning. Additionally, this enables us to complement the image-text alignment loss with a language modeling objective. On the MSCOCO test captions, our model establishes a new state of the art, outperforming its non-grounded version and, thus, confirming the effectiveness of visual groundings in constituency grammar induction. It also substantially outperforms the previous grounded model, with largest improvements on more `abstract' categories (e.g., +55.1% recall on VPs)." @default.
- W3100776995 created "2020-11-23" @default.
- W3100776995 creator A5057899933 @default.
- W3100776995 creator A5086717154 @default.
- W3100776995 date "2020-01-01" @default.
- W3100776995 modified "2023-09-24" @default.
- W3100776995 title "Visually Grounded Compound PCFGs" @default.
- W3100776995 cites W145849181 @default.
- W3100776995 cites W1495446613 @default.
- W3100776995 cites W1533861849 @default.
- W3100776995 cites W1599798800 @default.
- W3100776995 cites W1660460062 @default.
- W3100776995 cites W1848000237 @default.
- W3100776995 cites W1861492603 @default.
- W3100776995 cites W1970961429 @default.
- W3100776995 cites W1978470410 @default.
- W3100776995 cites W2047706513 @default.
- W3100776995 cites W2073132088 @default.
- W3100776995 cites W2081228205 @default.
- W3100776995 cites W2112184938 @default.
- W3100776995 cites W2113075298 @default.
- W3100776995 cites W2129882630 @default.
- W3100776995 cites W2133280805 @default.
- W3100776995 cites W2148398960 @default.
- W3100776995 cites W2150375089 @default.
- W3100776995 cites W2152907450 @default.
- W3100776995 cites W2156050092 @default.
- W3100776995 cites W2194775991 @default.
- W3100776995 cites W2250385622 @default.
- W3100776995 cites W2250646737 @default.
- W3100776995 cites W2250676463 @default.
- W3100776995 cites W2250742840 @default.
- W3100776995 cites W2252218513 @default.
- W3100776995 cites W2327562811 @default.
- W3100776995 cites W2911448865 @default.
- W3100776995 cites W2918073309 @default.
- W3100776995 cites W2932376173 @default.
- W3100776995 cites W2932637973 @default.
- W3100776995 cites W2949111363 @default.
- W3100776995 cites W2949399644 @default.
- W3100776995 cites W2949416428 @default.
- W3100776995 cites W2949579048 @default.
- W3100776995 cites W2963084773 @default.
- W3100776995 cites W2963264012 @default.
- W3100776995 cites W2963411763 @default.
- W3100776995 cites W2963651521 @default.
- W3100776995 cites W2963754491 @default.
- W3100776995 cites W2964030814 @default.
- W3100776995 cites W2964121744 @default.
- W3100776995 cites W2970488023 @default.
- W3100776995 cites W2995627237 @default.
- W3100776995 cites W3035720671 @default.
- W3100776995 cites W3037738964 @default.
- W3100776995 cites W3126921060 @default.
- W3100776995 doi "https://doi.org/10.18653/v1/2020.emnlp-main.354" @default.
- W3100776995 hasPublicationYear "2020" @default.
- W3100776995 type Work @default.
- W3100776995 sameAs 3100776995 @default.
- W3100776995 citedByCount "21" @default.
- W3100776995 countsByYear W31007769952021 @default.
- W3100776995 countsByYear W31007769952022 @default.
- W3100776995 countsByYear W31007769952023 @default.
- W3100776995 crossrefType "proceedings-article" @default.
- W3100776995 hasAuthorship W3100776995A5057899933 @default.
- W3100776995 hasAuthorship W3100776995A5086717154 @default.
- W3100776995 hasBestOaLocation W31007769951 @default.
- W3100776995 hasConcept C144024400 @default.
- W3100776995 hasConcept C156325361 @default.
- W3100776995 hasConcept C190248442 @default.
- W3100776995 hasConcept C36289849 @default.
- W3100776995 hasConcept C41008148 @default.
- W3100776995 hasConceptScore W3100776995C144024400 @default.
- W3100776995 hasConceptScore W3100776995C156325361 @default.
- W3100776995 hasConceptScore W3100776995C190248442 @default.
- W3100776995 hasConceptScore W3100776995C36289849 @default.
- W3100776995 hasConceptScore W3100776995C41008148 @default.
- W3100776995 hasLocation W31007769951 @default.
- W3100776995 hasLocation W31007769952 @default.
- W3100776995 hasOpenAccess W3100776995 @default.
- W3100776995 hasPrimaryLocation W31007769951 @default.
- W3100776995 hasRelatedWork W1511554945 @default.
- W3100776995 hasRelatedWork W1546533203 @default.
- W3100776995 hasRelatedWork W2015439768 @default.
- W3100776995 hasRelatedWork W2021436267 @default.
- W3100776995 hasRelatedWork W2620765995 @default.
- W3100776995 hasRelatedWork W2748952813 @default.
- W3100776995 hasRelatedWork W2899084033 @default.
- W3100776995 hasRelatedWork W2970776330 @default.
- W3100776995 hasRelatedWork W3040823075 @default.
- W3100776995 hasRelatedWork W3140115162 @default.
- W3100776995 isParatext "false" @default.
- W3100776995 isRetracted "false" @default.
- W3100776995 magId "3100776995" @default.
- W3100776995 workType "article" @default.