Matches in SemOpenAlex for { <https://semopenalex.org/work/W3040083890> ?p ?o ?g. }
- W3040083890 abstract "Stochastic gradient Langevin dynamics (SGLD) and stochastic gradient Hamiltonian Monte Carlo (SGHMC) are two popular Markov Chain Monte Carlo (MCMC) algorithms for Bayesian inference that can scale to large datasets, allowing to sample from the posterior distribution of the parameters of a statistical model given the input data and the prior distribution over the model parameters. However, these algorithms do not apply to the decentralized learning setting, when a network of agents are working collaboratively to learn the parameters of a statistical model without sharing their individual data due to privacy reasons or communication constraints. We study two algorithms: Decentralized SGLD (DE-SGLD) and Decentralized SGHMC (DE-SGHMC) which are adaptations of SGLD and SGHMC methods that allow scaleable Bayesian inference in the decentralized setting for large datasets. We show that when the posterior distribution is strongly log-concave and smooth, the iterates of these algorithms converge linearly to a neighborhood of the target distribution in the 2-Wasserstein distance if their parameters are selected appropriately. We illustrate the efficiency of our algorithms on decentralized Bayesian linear regression and Bayesian logistic regression problems." @default.
- W3040083890 created "2020-07-10" @default.
- W3040083890 creator A5006604791 @default.
- W3040083890 creator A5034370705 @default.
- W3040083890 creator A5035171822 @default.
- W3040083890 creator A5084475092 @default.
- W3040083890 date "2020-07-01" @default.
- W3040083890 modified "2023-09-22" @default.
- W3040083890 title "Decentralized Stochastic Gradient Langevin Dynamics and Hamiltonian Monte Carlo" @default.
- W3040083890 cites W123701566 @default.
- W3040083890 cites W1547358136 @default.
- W3040083890 cites W1568288633 @default.
- W3040083890 cites W1579303078 @default.
- W3040083890 cites W1616857247 @default.
- W3040083890 cites W1816220470 @default.
- W3040083890 cites W1865606400 @default.
- W3040083890 cites W1998754480 @default.
- W3040083890 cites W2040104067 @default.
- W3040083890 cites W2044212084 @default.
- W3040083890 cites W2101893470 @default.
- W3040083890 cites W2117905067 @default.
- W3040083890 cites W2122340819 @default.
- W3040083890 cites W2124768887 @default.
- W3040083890 cites W2128745767 @default.
- W3040083890 cites W2133482456 @default.
- W3040083890 cites W2137731592 @default.
- W3040083890 cites W2142819099 @default.
- W3040083890 cites W2144193737 @default.
- W3040083890 cites W2152766412 @default.
- W3040083890 cites W2165599843 @default.
- W3040083890 cites W2167433878 @default.
- W3040083890 cites W2167665486 @default.
- W3040083890 cites W2168354577 @default.
- W3040083890 cites W2168947765 @default.
- W3040083890 cites W2171911691 @default.
- W3040083890 cites W2186210550 @default.
- W3040083890 cites W2222154095 @default.
- W3040083890 cites W2338752163 @default.
- W3040083890 cites W2528062157 @default.
- W3040083890 cites W2552704617 @default.
- W3040083890 cites W2562674776 @default.
- W3040083890 cites W2570634592 @default.
- W3040083890 cites W2587943776 @default.
- W3040083890 cites W2592172378 @default.
- W3040083890 cites W2771477296 @default.
- W3040083890 cites W2772225710 @default.
- W3040083890 cites W2774971100 @default.
- W3040083890 cites W2799039767 @default.
- W3040083890 cites W2883697110 @default.
- W3040083890 cites W2883952565 @default.
- W3040083890 cites W2890773834 @default.
- W3040083890 cites W2892209803 @default.
- W3040083890 cites W2904091245 @default.
- W3040083890 cites W2911398476 @default.
- W3040083890 cites W2914852071 @default.
- W3040083890 cites W2946038967 @default.
- W3040083890 cites W2962703386 @default.
- W3040083890 cites W2962794482 @default.
- W3040083890 cites W2962839956 @default.
- W3040083890 cites W2962844775 @default.
- W3040083890 cites W2962947632 @default.
- W3040083890 cites W2963096809 @default.
- W3040083890 cites W2963153890 @default.
- W3040083890 cites W2963481221 @default.
- W3040083890 cites W2963505370 @default.
- W3040083890 cites W2963545724 @default.
- W3040083890 cites W2963599479 @default.
- W3040083890 cites W2964010828 @default.
- W3040083890 cites W2964066689 @default.
- W3040083890 cites W2964301034 @default.
- W3040083890 cites W2964311894 @default.
- W3040083890 cites W2966244755 @default.
- W3040083890 cites W2967190275 @default.
- W3040083890 cites W2970282196 @default.
- W3040083890 cites W2970628695 @default.
- W3040083890 cites W2971156569 @default.
- W3040083890 cites W2973303672 @default.
- W3040083890 cites W2978995744 @default.
- W3040083890 cites W2980919526 @default.
- W3040083890 cites W2985114067 @default.
- W3040083890 cites W2989361410 @default.
- W3040083890 cites W2997208636 @default.
- W3040083890 cites W3005629823 @default.
- W3040083890 cites W3008463948 @default.
- W3040083890 cites W3010627535 @default.
- W3040083890 cites W3012645498 @default.
- W3040083890 cites W3015771480 @default.
- W3040083890 cites W3022064003 @default.
- W3040083890 cites W3022621540 @default.
- W3040083890 cites W3037225843 @default.
- W3040083890 cites W3040426824 @default.
- W3040083890 cites W3042135496 @default.
- W3040083890 cites W3045239610 @default.
- W3040083890 cites W3046789028 @default.
- W3040083890 cites W3047579933 @default.
- W3040083890 cites W3048505912 @default.
- W3040083890 cites W3084298117 @default.
- W3040083890 cites W3087736320 @default.
- W3040083890 cites W3101853953 @default.
- W3040083890 cites W3102608064 @default.