Matches in SemOpenAlex for { <https://semopenalex.org/work/W3208144977> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3208144977 abstract "Federated Learning (FL) is a powerful technique for training a model on a server with data from several clients in a privacy-preserving manner. In FL, a server sends the model to every client, who then train the model locally and send it back to the server. The server aggregates the updated models and repeats the process for several rounds. FL incurs significant communication costs, in particular when transmitting the updated local models from the clients back to the server. Recently proposed algorithms quantize the model parameters to efficiently compress FL communication. These algorithms typically have a quantization level that controls the compression factor. We find that dynamic adaptations of the quantization level can boost compression without sacrificing model quality. First, we introduce a time-adaptive quantization algorithm that increases the quantization level as training progresses. Second, we introduce a client-adaptive quantization algorithm that assigns each individual client the optimal quantization level at every round. Finally, we combine both algorithms into DAdaQuant, the doubly-adaptive quantization algorithm. Our experiments show that DAdaQuant consistently improves client$rightarrow$server compression, outperforming the strongest non-adaptive baselines by up to $2.8times$." @default.
- W3208144977 created "2021-11-08" @default.
- W3208144977 creator A5040762162 @default.
- W3208144977 creator A5048483915 @default.
- W3208144977 creator A5071229843 @default.
- W3208144977 date "2021-10-31" @default.
- W3208144977 modified "2023-09-24" @default.
- W3208144977 title "DAdaQuant: Doubly-adaptive quantization for communication-efficient Federated Learning." @default.
- W3208144977 cites W2541884796 @default.
- W3208144977 cites W2744999500 @default.
- W3208144977 cites W2769644379 @default.
- W3208144977 cites W2889797931 @default.
- W3208144977 cites W2902113386 @default.
- W3208144977 cites W2914328083 @default.
- W3208144977 cites W2963092133 @default.
- W3208144977 cites W2963374099 @default.
- W3208144977 cites W2963709416 @default.
- W3208144977 cites W2970821029 @default.
- W3208144977 cites W2975043678 @default.
- W3208144977 cites W2977090839 @default.
- W3208144977 cites W3033664100 @default.
- W3208144977 cites W3036723732 @default.
- W3208144977 cites W3037913581 @default.
- W3208144977 cites W3046183292 @default.
- W3208144977 cites W3093436911 @default.
- W3208144977 cites W3098903812 @default.
- W3208144977 cites W3099584108 @default.
- W3208144977 cites W3113075536 @default.
- W3208144977 cites W3115851078 @default.
- W3208144977 cites W3162286130 @default.
- W3208144977 cites W3184192107 @default.
- W3208144977 hasPublicationYear "2021" @default.
- W3208144977 type Work @default.
- W3208144977 sameAs 3208144977 @default.
- W3208144977 citedByCount "0" @default.
- W3208144977 crossrefType "posted-content" @default.
- W3208144977 hasAuthorship W3208144977A5040762162 @default.
- W3208144977 hasAuthorship W3208144977A5048483915 @default.
- W3208144977 hasAuthorship W3208144977A5071229843 @default.
- W3208144977 hasConcept C11413529 @default.
- W3208144977 hasConcept C154945302 @default.
- W3208144977 hasConcept C199833920 @default.
- W3208144977 hasConcept C28855332 @default.
- W3208144977 hasConcept C2992525071 @default.
- W3208144977 hasConcept C41008148 @default.
- W3208144977 hasConcept C80444323 @default.
- W3208144977 hasConcept C93372532 @default.
- W3208144977 hasConceptScore W3208144977C11413529 @default.
- W3208144977 hasConceptScore W3208144977C154945302 @default.
- W3208144977 hasConceptScore W3208144977C199833920 @default.
- W3208144977 hasConceptScore W3208144977C28855332 @default.
- W3208144977 hasConceptScore W3208144977C2992525071 @default.
- W3208144977 hasConceptScore W3208144977C41008148 @default.
- W3208144977 hasConceptScore W3208144977C80444323 @default.
- W3208144977 hasConceptScore W3208144977C93372532 @default.
- W3208144977 hasLocation W32081449771 @default.
- W3208144977 hasOpenAccess W3208144977 @default.
- W3208144977 hasPrimaryLocation W32081449771 @default.
- W3208144977 hasRelatedWork W2007321238 @default.
- W3208144977 hasRelatedWork W2034955839 @default.
- W3208144977 hasRelatedWork W2306476809 @default.
- W3208144977 hasRelatedWork W2401175101 @default.
- W3208144977 hasRelatedWork W2529787296 @default.
- W3208144977 hasRelatedWork W2730552275 @default.
- W3208144977 hasRelatedWork W2753976758 @default.
- W3208144977 hasRelatedWork W2775827968 @default.
- W3208144977 hasRelatedWork W2963582525 @default.
- W3208144977 hasRelatedWork W2977090839 @default.
- W3208144977 hasRelatedWork W3039366294 @default.
- W3208144977 hasRelatedWork W3111857673 @default.
- W3208144977 hasRelatedWork W3133295491 @default.
- W3208144977 hasRelatedWork W3147053158 @default.
- W3208144977 hasRelatedWork W3157111505 @default.
- W3208144977 hasRelatedWork W3164411344 @default.
- W3208144977 hasRelatedWork W3189274515 @default.
- W3208144977 hasRelatedWork W3196370796 @default.
- W3208144977 hasRelatedWork W3204883510 @default.
- W3208144977 hasRelatedWork W192012345 @default.
- W3208144977 isParatext "false" @default.
- W3208144977 isRetracted "false" @default.
- W3208144977 magId "3208144977" @default.
- W3208144977 workType "article" @default.