In this paper, we focus on the family of Dual Coordinate Descent (DCD) algorithms that applies stochastic coordinate descent to solve the dual problem while ...
In this work, we propose a simple and computational efficient way to fix the convergence problem of PASSCoDe. Instead of allowing all worker threads to conduct ...
In theory, the algorithm converges only when the delay is small enough, but in practice the delay depends on the underlying parallel computing environment and ...
Additionally, we also need to control the gap between workers as update information from slow workers may lead to divergence [32] .
Fixing the convergence problems in parallel asynchronous dual ...
experts.illinois.edu › fingerprints
Dive into the research topics of 'Fixing the convergence problems in parallel asynchronous dual coordinate descent'. Together they form a unique fingerprint.
Stochastic Dual Coordinate Descent (DCD) is one of the most efficient ways to solve the family of L2-regularized empirical risk minimization problems, ...
Missing: Fixing | Show results with:Fixing
PASSCoDe-fix is based on the PASSCoDe (Parallel ASynchronous Stochastic dual Co-ordinate Descent) algorithm, and we aim to fix the divergence problem of ...
In this paper, we propose two algorithms – Adaptive Hybrid algorithm and Lazy-Sync algorithm, to overcome the convergence issues in parallel execution.
, Fixing the convergence problems in parallel asynchronous dual coordinate descent. ; Huan H Zhang ;. , DOI: 10.1109/ICDM.2016.171 ,. , Proceedings - 16th IEEE ...
Apr 6, 2015 · In this paper, we propose a family of asynchronous stochastic dual coordinate descent algorithms (ASDCD). Each thread repeatedly selects a random dual variable.