DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim, Jin-Hwa | ko |
dc.contributor.author | On, Kyoung-Woon | ko |
dc.contributor.author | Lim, Woosang | ko |
dc.contributor.author | Kim, Jeonghee | ko |
dc.contributor.author | Ha, Jung-Woo | ko |
dc.contributor.author | Zhang, Byoung-Tak | ko |
dc.date.accessioned | 2023-07-04T06:16:58Z | - |
dc.date.available | 2023-07-04T06:16:58Z | - |
dc.date.created | 2023-06-08 | - |
dc.date.issued | 2017-04 | - |
dc.identifier.citation | 5th International Conference on Learning Representations, ICLR 2017 | - |
dc.identifier.uri | http://hdl.handle.net/10203/310286 | - |
dc.description.abstract | Bilinear models provide rich representations compared with linear models. They have been applied in various visual tasks, such as object recognition, segmentation, and visual question-answering, to get state-of-the-art performances taking advantage of the expanded representations. However, bilinear representations tend to be high-dimensional, limiting the applicability to computationally complex tasks. We propose low-rank bilinear pooling using Hadamard product for an efficient attention mechanism of multimodal learning. We show that our model outperforms compact bilinear pooling in visual question-answering tasks with the state-of-the-art results on the VQA dataset, having a better parsimonious property. | - |
dc.language | English | - |
dc.publisher | International Conference on Learning Representations, ICLR | - |
dc.title | Hadamard product for low-rank bilinear pooling | - |
dc.type | Conference | - |
dc.identifier.scopusid | 2-s2.0-85087529518 | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | 5th International Conference on Learning Representations, ICLR 2017 | - |
dc.identifier.conferencecountry | FR | - |
dc.identifier.conferencelocation | Toulon | - |
dc.contributor.nonIdAuthor | Kim, Jin-Hwa | - |
dc.contributor.nonIdAuthor | On, Kyoung-Woon | - |
dc.contributor.nonIdAuthor | Kim, Jeonghee | - |
dc.contributor.nonIdAuthor | Ha, Jung-Woo | - |
dc.contributor.nonIdAuthor | Zhang, Byoung-Tak | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.