FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385216542> ?p ?o ?g. }

• W4385216542 endingPage "24" @default.
• W4385216542 startingPage "1" @default.
• W4385216542 abstract "The performance of current Scene Graph Generation (SGG) models is severely hampered by hard-to-distinguish predicates, e.g., woman-on/standing on/walking on-beach. As general SGG models tend to predict head predicates and re-balancing strategies prefer tail categories, none of them can appropriately handle hard-to-distinguish predicates. To tackle this issue, inspired by fine-grained image classification, which focuses on differentiating hard-to-distinguish objects, we propose an Adaptive Fine-Grained Predicates Learning (FGPL-A) which aims at differentiating hard-to-distinguish predicates for SGG. First, we introduce an Adaptive Predicate Lattice (PL-A) to figure out hard-to-distinguish predicates, which adaptively explores predicate correlations in keeping with model's dynamic learning pace. Practically, PL-A is initialized from SGG dataset, and gets refined by exploring model's predictions of current mini-batch. Utilizing PL-A, we propose an Adaptive Category Discriminating Loss (CDL-A) and an Adaptive Entity Discriminating Loss (EDL-A), which progressively regularize model's discriminating process with fine-grained supervision concerning model's dynamic learning status, ensuring balanced and efficient learning process. Extensive experimental results show that our proposed model-agnostic strategy significantly boosts performance of benchmark models on VG-SGG and GQA-SGG datasets by up to 175% and 76% on Mean Recall@100, achieving new state-of-the-art performance. Moreover, experiments on Sentence-to-Graph Retrieval and Image Captioning tasks further demonstrate practicability of our method." @default.
• W4385216542 created "2023-07-25" @default.
• W4385216542 creator A5021504282 @default.
• W4385216542 creator A5036987388 @default.
• W4385216542 creator A5066645546 @default.
• W4385216542 creator A5087623065 @default.
• W4385216542 creator A5090700061 @default.
• W4385216542 date "2023-01-01" @default.
• W4385216542 modified "2023-10-06" @default.
• W4385216542 title "Adaptive Fine-Grained Predicates Learning for Scene Graph Generation" @default.
• W4385216542 cites W1956340063 @default.
• W4385216542 cites W2133459682 @default.
• W4385216542 cites W2183182206 @default.
• W4385216542 cites W2506483933 @default.
• W4385216542 cites W2574790321 @default.
• W4385216542 cites W2579549467 @default.
• W4385216542 cites W2591644541 @default.
• W4385216542 cites W2605736949 @default.
• W4385216542 cites W2883170015 @default.
• W4385216542 cites W2886970679 @default.
• W4385216542 cites W2947385066 @default.
• W4385216542 cites W2948672349 @default.
• W4385216542 cites W2962737704 @default.
• W4385216542 cites W2962785943 @default.
• W4385216542 cites W2963101956 @default.
• W4385216542 cites W2963351448 @default.
• W4385216542 cites W2963393555 @default.
• W4385216542 cites W2963407932 @default.
• W4385216542 cites W2963514444 @default.
• W4385216542 cites W2963518342 @default.
• W4385216542 cites W2963536419 @default.
• W4385216542 cites W2963650529 @default.
• W4385216542 cites W2963691377 @default.
• W4385216542 cites W2963938081 @default.
• W4385216542 cites W2969679616 @default.
• W4385216542 cites W2980248933 @default.
• W4385216542 cites W2986821660 @default.
• W4385216542 cites W2990495699 @default.
• W4385216542 cites W2992049873 @default.
• W4385216542 cites W2992478697 @default.
• W4385216542 cites W2997426000 @default.
• W4385216542 cites W3008809756 @default.
• W4385216542 cites W3034538190 @default.
• W4385216542 cites W3034933032 @default.
• W4385216542 cites W3035017890 @default.
• W4385216542 cites W3035552357 @default.
• W4385216542 cites W3036869132 @default.
• W4385216542 cites W3106328333 @default.
• W4385216542 cites W3108864070 @default.
• W4385216542 cites W3166228617 @default.
• W4385216542 cites W3174293538 @default.
• W4385216542 cites W3174865181 @default.
• W4385216542 cites W3176474016 @default.
• W4385216542 cites W3180157652 @default.
• W4385216542 cites W3181556077 @default.
• W4385216542 cites W3182635745 @default.
• W4385216542 cites W3182902595 @default.
• W4385216542 cites W3191092153 @default.
• W4385216542 cites W3196936439 @default.
• W4385216542 cites W3203170887 @default.
• W4385216542 cites W3205051885 @default.
• W4385216542 cites W3215287157 @default.
• W4385216542 cites W4214685694 @default.
• W4385216542 cites W4214693531 @default.
• W4385216542 cites W4221150632 @default.
• W4385216542 cites W4225868495 @default.
• W4385216542 cites W4225998785 @default.
• W4385216542 cites W4283463548 @default.
• W4385216542 cites W4285287023 @default.
• W4385216542 cites W4293518226 @default.
• W4385216542 cites W4312361652 @default.
• W4385216542 cites W4312429751 @default.
• W4385216542 cites W4312555984 @default.
• W4385216542 cites W4312563197 @default.
• W4385216542 cites W4312578903 @default.
• W4385216542 cites W4312682661 @default.
• W4385216542 cites W4312744661 @default.
• W4385216542 cites W4313037583 @default.
• W4385216542 cites W4313161463 @default.
• W4385216542 cites W639708223 @default.
• W4385216542 doi "https://doi.org/10.1109/tpami.2023.3298356" @default.
• W4385216542 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37788219" @default.
• W4385216542 hasPublicationYear "2023" @default.
• W4385216542 type Work @default.
• W4385216542 citedByCount "0" @default.
• W4385216542 crossrefType "journal-article" @default.
• W4385216542 hasAuthorship W4385216542A5021504282 @default.
• W4385216542 hasAuthorship W4385216542A5036987388 @default.
• W4385216542 hasAuthorship W4385216542A5066645546 @default.
• W4385216542 hasAuthorship W4385216542A5087623065 @default.
• W4385216542 hasAuthorship W4385216542A5090700061 @default.
• W4385216542 hasConcept C119857082 @default.
• W4385216542 hasConcept C132525143 @default.
• W4385216542 hasConcept C13280743 @default.
• W4385216542 hasConcept C140146324 @default.
• W4385216542 hasConcept C154945302 @default.
• W4385216542 hasConcept C179372163 @default.
• W4385216542 hasConcept C185798385 @default.

• next