Heterogeneous Multi-Party Learning With Data-Driven Network Sampling

Heterogeneous Multi-Party Learning With Data-Driven Network Sampling

Multi-party learning provides an effective approach for training a machine learning model, e.g., deep neural networks (DNNs), over decentralized data by leveraging multiple decentralized computing devices, subjected to legal and practical constraints. Different parties, so-called local participants,...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence. - 1979. - 45(2023), 11 vom: 28. Nov., Seite 13328-13343
1. Verfasser: Gong, Maoguo (VerfasserIn)
Weitere Verfasser: Gao, Yuan, Wu, Yue, Zhang, Yuanqiao, Qin, A K, Ong, Yew-Soon
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2023
Zugriff auf das übergeordnete Werk:IEEE transactions on pattern analysis and machine intelligence
Schlagworte:Journal Article
LEADER 01000naa a22002652 4500
001 NLM358804981
003 DE-627
005 20231226075523.0
007 cr uuu---uuuuu
008 231226s2023 xx |||||o 00| ||eng c
024 7 |a 10.1109/TPAMI.2023.3290213  |2 doi 
028 5 2 |a pubmed24n1195.xml 
035 |a (DE-627)NLM358804981 
035 |a (NLM)37379198 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Gong, Maoguo  |e verfasserin  |4 aut 
245 1 0 |a Heterogeneous Multi-Party Learning With Data-Driven Network Sampling 
264 1 |c 2023 
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 04.10.2023 
500 |a published: Print-Electronic 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Multi-party learning provides an effective approach for training a machine learning model, e.g., deep neural networks (DNNs), over decentralized data by leveraging multiple decentralized computing devices, subjected to legal and practical constraints. Different parties, so-called local participants, usually provide heterogenous data in a decentralized mode, leading to non-IID data distributions across different local participants which pose a notorious challenge for multi-party learning. To address this challenge, we propose a novel heterogeneous differentiable sampling (HDS) framework. Inspired by the dropout strategy in DNNs, a data-driven network sampling strategy is devised in the HDS framework, with differentiable sampling rates which allow each local participant to extract from a common global model the optimal local model that best adapts to its own data properties so that the size of the local model can be significantly reduced to enable more efficient inference. Meanwhile, co-adaptation of the global model via learning such local models allows for achieving better learning performance under non-IID data distributions and speeds up the convergence of the global model. Experiments have demonstrated the superiority of the proposed method over several popular multi-party learning techniques in the multi-party settings with non-IID data distributions 
650 4 |a Journal Article 
700 1 |a Gao, Yuan  |e verfasserin  |4 aut 
700 1 |a Wu, Yue  |e verfasserin  |4 aut 
700 1 |a Zhang, Yuanqiao  |e verfasserin  |4 aut 
700 1 |a Qin, A K  |e verfasserin  |4 aut 
700 1 |a Ong, Yew-Soon  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on pattern analysis and machine intelligence  |d 1979  |g 45(2023), 11 vom: 28. Nov., Seite 13328-13343  |w (DE-627)NLM098212257  |x 1939-3539  |7 nnns 
773 1 8 |g volume:45  |g year:2023  |g number:11  |g day:28  |g month:11  |g pages:13328-13343 
856 4 0 |u http://dx.doi.org/10.1109/TPAMI.2023.3290213  |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 45  |j 2023  |e 11  |b 28  |c 11  |h 13328-13343