Compressed Decentralized Momentum Stochastic Gradient Methods for Nonconvex Optimization
Wei Liu, Anweshit Panda, Ujwal Pandey et al.
Action editor: Anastasios Kyrillidis
https://openreview.net/forum?id=RqhMQHHkB4
#compression #compressed #nonconvex
Adaptive Gradient Normalization and Independent Sampling for (Stochastic) Generalized-Smooth Opti...
Yufeng Yang, Erin E. Tripp, Yifan Sun, Shaofeng Zou, Yi Zhou
Action editor: Sebastian Stich
https://openreview.net/forum?id=KKSQQMlEfw
#normalization #gradient #nonconvex
Compressed Decentralized Momentum Stochastic Gradient Methods for Nonconvex Optimization
New #TMLR-Paper-with-Video:
Compressed Decentralized Momentum Stochastic Gradient Methods for Nonconvex Optimization
Wei Liu, Anweshit Panda, Ujwal Pandey et al.
https://tmlr.infinite-conf.org/paper_pages/RqhMQHHkB4
#compression #compressed #nonconvex
Zeroth-Order Methods Advance Stochastic Nonconvex Optimization
Two zeroth‑order algorithms were introduced to solve stochastic nonconvex nonsmooth composite optimization without smoothness assumptions, providing finite convergence guarantees (6 Oct 2025). getnews.me/zeroth-order-methods-adv... #zerothorder #nonconvex
Double‑Optimism Method Boosts Online‑to‑Nonconvex Optimization
Researchers introduced a double‑optimism method that eliminates the double‑loop and removes the log factor, achieving O(ε⁻¹·⁷⁵) and unified O(ε⁻¹·⁷⁵ + σ² ε⁻³·⁵) complexities. Read more: getnews.me/double-optimism-method-b... #nonconvex #optimization
Decentralized Stochastic Gradient Method Advances Nonconvex Optimization
DNSGD finds ε‑stationary points in nonconvex optimization, cutting sample complexity per agent as agents increase; communication rounds scale with network’s spectral gap. Read more: getnews.me/decentralized-stochastic... #decentralized #nonconvex
Nonconvex Regularization Boosts Feature Selection in RL
Batch algorithm uses a PMC penalty and FRBS optimization to prune RL features, dropping irrelevant variables while preserving policy performance on benchmarks. Read more: getnews.me/nonconvex-regularization... #reinforcementlearning #featureselection #nonconvex
Preconditioned Third‑Order IMEX Algorithm for Non‑Convex Optimization
A IMEX algorithm using Adams‑Bashforth and BDF steps converges by the Kurdyka–Łojasiewicz property and outperforms DC methods on SCAD‑regularized least‑squares. Read more: getnews.me/preconditioned-third-ord... #imex #nonconvex
