FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2904658088> ?p ?o ?g. }

• W2904658088 abstract "While significant progress has been made in the image captioning task, video description is still in its infancy due to the complex nature of video data. Generating multi-sentence descriptions for long videos is even more challenging. Among the main issues are the fluency and coherence of the generated descriptions, and their relevance to the video. Recently, reinforcement and adversarial learning based methods have been explored to improve the image captioning models; however, both types of methods suffer from a number of issues, e.g. poor readability and high redundancy for RL and stability issues for GANs. In this work, we instead propose to apply adversarial techniques during inference, designing a discriminator which encourages better multi-sentence video description. In addition, we find that a multi-discriminator hybrid design, where each discriminator targets one aspect of a description, leads to the best results. Specifically, we decouple the discriminator to evaluate on three criteria: 1) visual relevance to the video, 2) language diversity and fluency, and 3) coherence across sentences. Our approach results in more accurate, diverse, and coherent multi-sentence video descriptions, as shown by automatic as well as human evaluation on the popular ActivityNet Captions dataset." @default.
• W2904658088 created "2018-12-22" @default.
• W2904658088 creator A5024481540 @default.
• W2904658088 creator A5029105520 @default.
• W2904658088 creator A5030680279 @default.
• W2904658088 creator A5037747070 @default.
• W2904658088 date "2018-12-13" @default.
• W2904658088 modified "2023-10-01" @default.
• W2904658088 title "Adversarial Inference for Multi-Sentence Video Description" @default.
• W2904658088 cites W1522301498 @default.
• W2904658088 cites W1596841185 @default.
• W2904658088 cites W1601567445 @default.
• W2904658088 cites W1893116441 @default.
• W2904658088 cites W1947481528 @default.
• W2904658088 cites W1956340063 @default.
• W2904658088 cites W2064675550 @default.
• W2904658088 cites W2099471712 @default.
• W2904658088 cites W2101105183 @default.
• W2904658088 cites W2108598243 @default.
• W2904658088 cites W2119717200 @default.
• W2904658088 cites W2133459682 @default.
• W2904658088 cites W2156737235 @default.
• W2904658088 cites W2194775991 @default.
• W2904658088 cites W2250539671 @default.
• W2904658088 cites W2277195237 @default.
• W2904658088 cites W2337353209 @default.
• W2904658088 cites W2405676915 @default.
• W2904658088 cites W2507009361 @default.
• W2904658088 cites W2508429489 @default.
• W2904658088 cites W2547875792 @default.
• W2904658088 cites W2557414982 @default.
• W2904658088 cites W2565656701 @default.
• W2904658088 cites W2593383075 @default.
• W2904658088 cites W2597985671 @default.
• W2904658088 cites W2610163825 @default.
• W2904658088 cites W2619947201 @default.
• W2904658088 cites W2620623908 @default.
• W2904658088 cites W2729842244 @default.
• W2904658088 cites W2742943414 @default.
• W2904658088 cites W2745461083 @default.
• W2904658088 cites W2768287968 @default.
• W2904658088 cites W2784025607 @default.
• W2904658088 cites W2790888757 @default.
• W2904658088 cites W2794360013 @default.
• W2904658088 cites W2795215224 @default.
• W2904658088 cites W2795840542 @default.
• W2904658088 cites W2796207103 @default.
• W2904658088 cites W2796239628 @default.
• W2904658088 cites W2798490859 @default.
• W2904658088 cites W2798793675 @default.
• W2904658088 cites W2799042952 @default.
• W2904658088 cites W2799047197 @default.
• W2904658088 cites W2883512601 @default.
• W2904658088 cites W2886748926 @default.
• W2904658088 cites W2887097088 @default.
• W2904658088 cites W2891939431 @default.
• W2904658088 cites W2892087592 @default.
• W2904658088 cites W2900260828 @default.
• W2904658088 cites W2948648140 @default.
• W2904658088 cites W2949197413 @default.
• W2904658088 cites W2949954138 @default.
• W2904658088 cites W2949999304 @default.
• W2904658088 cites W2950019618 @default.
• W2904658088 cites W2950307714 @default.
• W2904658088 cites W2950401034 @default.
• W2904658088 cites W2950438040 @default.
• W2904658088 cites W2950672580 @default.
• W2904658088 cites W2951569546 @default.
• W2904658088 cites W2951684117 @default.
• W2904658088 cites W2952591111 @default.
• W2904658088 cites W2953106684 @default.
• W2904658088 cites W2953137234 @default.
• W2904658088 cites W2962793481 @default.
• W2904658088 cites W2962934715 @default.
• W2904658088 cites W2963033554 @default.
• W2904658088 cites W2963073614 @default.
• W2904658088 cites W2963084599 @default.
• W2904658088 cites W2963177403 @default.
• W2904658088 cites W2963248296 @default.
• W2904658088 cites W2963283805 @default.
• W2904658088 cites W2963576560 @default.
• W2904658088 cites W2963843052 @default.
• W2904658088 cites W2964268978 @default.
• W2904658088 hasPublicationYear "2018" @default.
• W2904658088 type Work @default.
• W2904658088 sameAs 2904658088 @default.
• W2904658088 citedByCount "0" @default.
• W2904658088 crossrefType "posted-content" @default.
• W2904658088 hasAuthorship W2904658088A5024481540 @default.
• W2904658088 hasAuthorship W2904658088A5029105520 @default.
• W2904658088 hasAuthorship W2904658088A5030680279 @default.
• W2904658088 hasAuthorship W2904658088A5037747070 @default.
• W2904658088 hasConcept C111919701 @default.
• W2904658088 hasConcept C115961682 @default.
• W2904658088 hasConcept C119857082 @default.
• W2904658088 hasConcept C138885662 @default.
• W2904658088 hasConcept C152124472 @default.
• W2904658088 hasConcept C154945302 @default.
• W2904658088 hasConcept C157657479 @default.
• W2904658088 hasConcept C158154518 @default.

• next