Mathematics – Optimization and Control
Scientific paper
2011-05-24
Mathematics
Optimization and Control
36 pages, 2 figures
Scientific paper
We generalize stochastic subgradient descent methods to situations in which we do not receive independent samples from the distribution over which we optimize, but instead receive samples that are coupled over time. We show that as long as the source of randomness is suitably ergodic---it converges quickly enough to a stationary distribution---the method enjoys strong convergence guarantees, both in expectation and with high probability. This result has implications for stochastic optimization in high-dimensional spaces, peer-to-peer distributed optimization schemes, decision problems with dependent data, and stochastic optimization problems over combinatorial spaces.
Agarwal Alekh
Duchi John C.
Johansson Mikael
Jordan Michael I.
No associations
LandOfFree
Ergodic Subgradient Descent does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Ergodic Subgradient Descent, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ergodic Subgradient Descent will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-515285