Matches in SemOpenAlex for { <https://semopenalex.org/work/W2996974038> ?p ?o ?g. }
- W2996974038 abstract "Federated learning is a method of training models on private data distributed over multiple devices. To keep device data private, the global model is trained by only communicating parameters and updates which poses scalability challenges for large models. To this end, we propose a new federated learning algorithm that jointly learns compact local representations on each device and a global model across all devices. As a result, the global model can be smaller since it only operates on local representations, reducing the number of communicated parameters. Theoretically, we provide a generalization analysis which shows that a combination of local and global models reduces both variance in the data as well as variance across device distributions. Empirically, we demonstrate that local models enable communication-efficient training while retaining performance. We also evaluate on the task of personalized mood prediction from real-world mobile data where privacy is key. Finally, local models handle heterogeneous data from new devices, and learn fair representations that obfuscate protected attributes such as race, age, and gender." @default.
- W2996974038 created "2020-01-10" @default.
- W2996974038 creator A5004510687 @default.
- W2996974038 creator A5071983998 @default.
- W2996974038 creator A5076016952 @default.
- W2996974038 creator A5081398601 @default.
- W2996974038 creator A5086233510 @default.
- W2996974038 date "2020-01-06" @default.
- W2996974038 modified "2023-10-18" @default.
- W2996974038 title "Think Locally, Act Globally: Federated Learning with Local and Global Representations" @default.
- W2996974038 cites W1559060276 @default.
- W2996974038 cites W1933349210 @default.
- W2996974038 cites W1964862779 @default.
- W2996974038 cites W2027595342 @default.
- W2996974038 cites W2064675550 @default.
- W2996974038 cites W2099471712 @default.
- W2996974038 cites W2112796928 @default.
- W2996974038 cites W2143104527 @default.
- W2996974038 cites W2162670686 @default.
- W2996974038 cites W2172166488 @default.
- W2996974038 cites W2194775991 @default.
- W2996974038 cites W2321533354 @default.
- W2996974038 cites W2324203673 @default.
- W2996974038 cites W2523153599 @default.
- W2996974038 cites W2541884796 @default.
- W2996974038 cites W2546563948 @default.
- W2996974038 cites W2547352193 @default.
- W2996974038 cites W2554980225 @default.
- W2996974038 cites W2556013083 @default.
- W2996974038 cites W2610074929 @default.
- W2996974038 cites W2619383789 @default.
- W2996974038 cites W2753338564 @default.
- W2996974038 cites W2785456003 @default.
- W2996974038 cites W2787752464 @default.
- W2996974038 cites W2788193959 @default.
- W2996974038 cites W2797532987 @default.
- W2996974038 cites W2799258637 @default.
- W2996974038 cites W2888161220 @default.
- W2996974038 cites W2900594532 @default.
- W2996974038 cites W2902113386 @default.
- W2996974038 cites W2905444686 @default.
- W2996974038 cites W2908866177 @default.
- W2996974038 cites W2911457688 @default.
- W2996974038 cites W2911978475 @default.
- W2996974038 cites W2912934387 @default.
- W2996974038 cites W2913010492 @default.
- W2996974038 cites W2913570153 @default.
- W2996974038 cites W2914328083 @default.
- W2996974038 cites W2917462349 @default.
- W2996974038 cites W2922267690 @default.
- W2996974038 cites W2922862135 @default.
- W2996974038 cites W2923764619 @default.
- W2996974038 cites W2943618673 @default.
- W2996974038 cites W2949505205 @default.
- W2996974038 cites W2950018712 @default.
- W2996974038 cites W2962688627 @default.
- W2996974038 cites W2962835968 @default.
- W2996974038 cites W2963300197 @default.
- W2996974038 cites W2963317470 @default.
- W2996974038 cites W2963803379 @default.
- W2996974038 cites W2963813679 @default.
- W2996974038 cites W2964166170 @default.
- W2996974038 cites W2964256806 @default.
- W2996974038 cites W2967536008 @default.
- W2996974038 cites W2971092532 @default.
- W2996974038 cites W3006035290 @default.
- W2996974038 cites W3010262580 @default.
- W2996974038 cites W3035591180 @default.
- W2996974038 hasPublicationYear "2020" @default.
- W2996974038 type Work @default.
- W2996974038 sameAs 2996974038 @default.
- W2996974038 citedByCount "64" @default.
- W2996974038 countsByYear W29969740382020 @default.
- W2996974038 countsByYear W29969740382021 @default.
- W2996974038 countsByYear W29969740382022 @default.
- W2996974038 crossrefType "posted-content" @default.
- W2996974038 hasAuthorship W2996974038A5004510687 @default.
- W2996974038 hasAuthorship W2996974038A5071983998 @default.
- W2996974038 hasAuthorship W2996974038A5076016952 @default.
- W2996974038 hasAuthorship W2996974038A5081398601 @default.
- W2996974038 hasAuthorship W2996974038A5086233510 @default.
- W2996974038 hasConcept C119857082 @default.
- W2996974038 hasConcept C121955636 @default.
- W2996974038 hasConcept C134306372 @default.
- W2996974038 hasConcept C136764020 @default.
- W2996974038 hasConcept C144133560 @default.
- W2996974038 hasConcept C154945302 @default.
- W2996974038 hasConcept C162324750 @default.
- W2996974038 hasConcept C177148314 @default.
- W2996974038 hasConcept C186967261 @default.
- W2996974038 hasConcept C187736073 @default.
- W2996974038 hasConcept C196083921 @default.
- W2996974038 hasConcept C26517878 @default.
- W2996974038 hasConcept C2780451532 @default.
- W2996974038 hasConcept C2992525071 @default.
- W2996974038 hasConcept C33923547 @default.
- W2996974038 hasConcept C38652104 @default.
- W2996974038 hasConcept C41008148 @default.
- W2996974038 hasConcept C48044578 @default.
- W2996974038 hasConcept C77088390 @default.