Presentation | 2019-06-17 Reliability Assessment by Bayesian Deep Learning for Image-Caption Retrieval Task Kenta Hama, Takashi Matsubara, Kuniaki Uehara, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Following the development of black-box machine learning algorithms, the practical demand of the re- liability assessment is rapidly rising. Recent progress in Bayesian deep learning has enabled us to quantify the uncertainty of its output, potentially providing a reliability measure. While many previous studies have evaluated the uncertainty measures for classification and regression tasks, their approaches are not always applicable to other tasks. This study investigates two sides of image-caption embedding and retrieval systems The embedding task is similar to the regression task, and the model averaging based on the regression improves the retrieval performance. However, its uncertainty measure cannot evaluate the reliability of retrieval appropriately, and the uncertainty mea- sure for the classification task is applicable. This study confirms that this tendency is common among datasets, DNN architectures, and similarity functions. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | multi-modal embedding / image-caption retrieval / uncertainty quantification |
Paper # | IBISML2019-1 |
Date of Issue | 2019-06-10 (IBISML) |
Conference Information | |
Committee | NC / IBISML / IPSJ-MPS / IPSJ-BIO |
---|---|
Conference Date | 2019/6/17(3days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Okinawa Institute of Science and Technology |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Neurocomputing, Machine Learning Approach to Biodata Mining, and General |
Chair | Hayaru Shouno(UEC) / Hisashi Kashima(Kyoto Univ.) / Masakazu Sekijima(Tokyo Tech) / Hiroyuki Kurata(Kyutech) |
Vice Chair | Kazuyuki Samejima(Tamagawa Univ) / Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo) |
Secretary | Kazuyuki Samejima(NAIST) / Masashi Sugiyama(NTT) / Koji Tsuda(Nagoya Inst. of Tech.) / (AIST) / (Nagoya Univ.) |
Assistant | Takashi Shinozaki(NICT) / Ken Takiyama(TUAT) / Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.) |
Paper Information | |
Registration To | Technical Committee on Neurocomputing / Technical Committee on Infomation-Based Induction Sciences and Machine Learning / IPSJ Special Interest Group on Mathematical Modeling and Problem Solving / IPSJ Special Interest Group on Bioinformatics and Genomics |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Reliability Assessment by Bayesian Deep Learning for Image-Caption Retrieval Task |
Sub Title (in English) | |
Keyword(1) | multi-modal embedding |
Keyword(2) | image-caption retrieval |
Keyword(3) | uncertainty quantification |
1st Author's Name | Kenta Hama |
1st Author's Affiliation | Kobe University(Kobe Univ.) |
2nd Author's Name | Takashi Matsubara |
2nd Author's Affiliation | Kobe University(Kobe Univ.) |
3rd Author's Name | Kuniaki Uehara |
3rd Author's Affiliation | Kobe University(Kobe Univ.) |
Date | 2019-06-17 |
Paper # | IBISML2019-1 |
Volume (vol) | vol.119 |
Number (no) | IBISML-89 |
Page | pp.pp.1-8(IBISML), |
#Pages | 8 |
Date of Issue | 2019-06-10 (IBISML) |