On Stochastic Gradient Langevin Dynamics with Dependent Data Streams: The Fully Nonconvex Case

DOI10.1137/20M1355392MaRDI QIDQ5162623

Miklós Rásonyi, Sotirios Sabanis, Ying Zhang, N. H. Chau, Eric Moulines

Publication date: 3 November 2021

Published in: SIAM Journal on Mathematics of Data Science (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1905.13142

zbMATH Keywords

contractionnonconvex optimizationLangevin dynamicsstochastic gradientconvergence guarantees

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Stochastic learning and adaptive control (93E35) Sequential statistical analysis (62L10)

Related Items

Stochastic gradient Hamiltonian Monte Carlo for non-convex learning, Full error analysis for the training of deep neural networks, Global Optimization via Schrödinger–Föllmer Diffusion, Functional central limit theorem and strong law of large numbers for stochastic gradient Langevin dynamics, Optimising portfolio diversification and dimensionality, Nonasymptotic estimates for stochastic gradient Langevin dynamics under local conditions in nonconvex optimization, Distributed event-triggered unadjusted Langevin algorithm for Bayesian learning, Convergence of the Kiefer–Wolfowitz algorithm in the presence of discontinuities, Taming Neural Networks with TUSLA: Nonconvex Learning via Adaptive Stochastic Gradient Langevin Algorithms, Multi-index antithetic stochastic gradient algorithm, Unnamed Item, On the stability of the stochastic gradient Langevin algorithm with dependent data stream, Unadjusted Langevin algorithm for sampling a mixture of weakly smooth potentials

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Forecasting the term structure of government bond yields
• Sparse regression learning by aggregation and Langevin Monte-Carlo
• On stochastic gradient Langevin dynamics with dependent data streams in the logconcave case
• The tamed unadjusted Langevin algorithm
• User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
• High-dimensional Bayesian inference via the unadjusted Langevin algorithm
• Higher order Langevin Monte Carlo algorithm
• Couplings and quantitative contraction rates for Langevin dynamics
• Optimal posting price of limit orders: learning by trading
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• Nonasymptotic bounds for sampling algorithms without log-concavity
• Stochastic approximation with averaging innovation applied to Finance
• On a class of mixing processes
• Volatility is rough
• Markov Chains
• Deep learning for finance: deep portfolios
• Quantitative Harris-type theorems for diffusions and McKean–Vlasov processes
• Empirical properties of asset returns: stylized facts and statistical issues
• On fixed gain recursive estimators with discontinuity in the parameters
• Machine Learning in Finance
• Optimal Split of Orders Across Liquidity Pools: A Stochastic Algorithm Approach
• Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
• Optimal Transport