Transferable Representation Learning with Deep Adaptation Networks

Transferable Representation Learning with Deep Adaptation Networks

Domain adaptation studies learning algorithms that generalize across source domains and target domains that exhibit different distributions. Recent studies reveal that deep neural networks can learn transferable features that generalize well to similar novel tasks. However, as deep features eventual...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence. - 1979. - 41(2019), 12 vom: 01. Dez., Seite 3071-3085
1. Verfasser: Long, Mingsheng (VerfasserIn)
Weitere Verfasser: Cao, Yue, Cao, Zhangjie, Wang, Jianmin, Jordan, Michael I
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2019
Zugriff auf das übergeordnete Werk:IEEE transactions on pattern analysis and machine intelligence
Schlagworte:Journal Article
LEADER 01000naa a22002652 4500
001 NLM288262824
003 DE-627
005 20231225055857.0
007 cr uuu---uuuuu
008 231225s2019 xx |||||o 00| ||eng c
024 7 |a 10.1109/TPAMI.2018.2868685  |2 doi 
028 5 2 |a pubmed24n0960.xml 
035 |a (DE-627)NLM288262824 
035 |a (NLM)30188813 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Long, Mingsheng  |e verfasserin  |4 aut 
245 1 0 |a Transferable Representation Learning with Deep Adaptation Networks 
264 1 |c 2019 
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 Revised 04.03.2020 
500 |a published: Print-Electronic 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Domain adaptation studies learning algorithms that generalize across source domains and target domains that exhibit different distributions. Recent studies reveal that deep neural networks can learn transferable features that generalize well to similar novel tasks. However, as deep features eventually transition from general to specific along the network, feature transferability drops significantly in higher task-specific layers with increasing domain discrepancy. To formally reduce the effects of this discrepancy and enhance feature transferability in task-specific layers, we develop a novel framework for deep adaptation networks that extends deep convolutional neural networks to domain adaptation problems. The framework embeds the deep features of all task-specific layers into reproducing kernel Hilbert spaces (RKHSs) and optimally matches different domain distributions. The deep features are made more transferable by exploiting low-density separation of target-unlabeled data in very deep architectures, while the domain discrepancy is further reduced via the use of multiple kernel learning that enhances the statistical power of kernel embedding matching. The overall framework is cast in a minimax game setting. Extensive empirical evidence shows that the proposed networks yield state-of-the-art results on standard visual domain-adaptation benchmarks 
650 4 |a Journal Article 
700 1 |a Cao, Yue  |e verfasserin  |4 aut 
700 1 |a Cao, Zhangjie  |e verfasserin  |4 aut 
700 1 |a Wang, Jianmin  |e verfasserin  |4 aut 
700 1 |a Jordan, Michael I  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on pattern analysis and machine intelligence  |d 1979  |g 41(2019), 12 vom: 01. Dez., Seite 3071-3085  |w (DE-627)NLM098212257  |x 1939-3539  |7 nnns 
773 1 8 |g volume:41  |g year:2019  |g number:12  |g day:01  |g month:12  |g pages:3071-3085 
856 4 0 |u http://dx.doi.org/10.1109/TPAMI.2018.2868685  |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 41  |j 2019  |e 12  |b 01  |c 12  |h 3071-3085