Harmonized Multimodal Learning with Gaussian Process Latent Variable Models

Harmonized Multimodal Learning with Gaussian Process Latent Variable Models

Multimodal learning aims to discover the relationship between multiple modalities. It has become an important research topic due to extensive multimodal applications such as cross-modal retrieval. This paper attempts to address the modality heterogeneity problem based on Gaussian process latent vari...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence. - 1979. - 43(2021), 3 vom: 29. März, Seite 858-872
1. Verfasser: Song, Guoli (VerfasserIn)
Weitere Verfasser: Wang, Shuhui, Huang, Qingming, Tian, Qi
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2021
Zugriff auf das übergeordnete Werk:IEEE transactions on pattern analysis and machine intelligence
Schlagworte:Journal Article Research Support, Non-U.S. Gov't
LEADER 01000naa a22002652 4500
001 NLM301536511
003 DE-627
005 20231225104833.0
007 cr uuu---uuuuu
008 231225s2021 xx |||||o 00| ||eng c
024 7 |a 10.1109/TPAMI.2019.2942028  |2 doi 
028 5 2 |a pubmed24n1005.xml 
035 |a (DE-627)NLM301536511 
035 |a (NLM)31545710 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Song, Guoli  |e verfasserin  |4 aut 
245 1 0 |a Harmonized Multimodal Learning with Gaussian Process Latent Variable Models 
264 1 |c 2021 
336 |a Text  |b txt  |2 rdacontent 
337 |a ƒaComputermedien  |b c  |2 rdamedia 
338 |a ƒa Online-Ressource  |b cr  |2 rdacarrier 
500 |a Date Completed 29.09.2021 
500 |a Date Revised 29.09.2021 
500 |a published: Print-Electronic 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Multimodal learning aims to discover the relationship between multiple modalities. It has become an important research topic due to extensive multimodal applications such as cross-modal retrieval. This paper attempts to address the modality heterogeneity problem based on Gaussian process latent variable models (GPLVMs) to represent multimodal data in a common space. Previous multimodal GPLVM extensions generally adopt individual learning schemes on latent representations and kernel hyperparameters, which ignore their intrinsic relationship. To exploit strong complementarity among different modalities and GPLVM components, we develop a novel learning scheme called Harmonization, where latent representations and kernel hyperparameters are jointly learned from each other. Beyond the correlation fitting or intra-modal structure preservation paradigms widely used in existing studies, the harmonization is derived in a model-driven manner to encourage the agreement between modality-specific GP kernels and the similarity of latent representations. We present a range of multimodal learning models by incorporating the harmonization mechanism into several representative GPLVM-based approaches. Experimental results on four benchmark datasets show that the proposed models outperform the strong baselines for cross-modal retrieval tasks, and that the harmonized multimodal learning method is superior in discovering semantically consistent latent representation 
650 4 |a Journal Article 
650 4 |a Research Support, Non-U.S. Gov't 
700 1 |a Wang, Shuhui  |e verfasserin  |4 aut 
700 1 |a Huang, Qingming  |e verfasserin  |4 aut 
700 1 |a Tian, Qi  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on pattern analysis and machine intelligence  |d 1979  |g 43(2021), 3 vom: 29. März, Seite 858-872  |w (DE-627)NLM098212257  |x 1939-3539  |7 nnns 
773 1 8 |g volume:43  |g year:2021  |g number:3  |g day:29  |g month:03  |g pages:858-872 
856 4 0 |u http://dx.doi.org/10.1109/TPAMI.2019.2942028  |3 Volltext 
912 |a GBV_USEFLAG_A 
912 |a SYSFLAG_A 
912 |a GBV_NLM 
912 |a GBV_ILN_350 
951 |a AR 
952 |d 43  |j 2021  |e 3  |b 29  |c 03  |h 858-872