Login / Signup
A Markov Chain Theory Approach to Characterizing the Minimax Optimality of Stochastic Gradient Descent (for Least Squares).
Prateek Jain
Sham M. Kakade
Rahul Kidambi
Praneeth Netrapalli
Venkata Krishna Pillutla
Aaron Sidford
Published in:
FSTTCS (2017)
Keyphrases
</>
markov chain
least squares
stochastic gradient descent
importance sampling
steady state
monte carlo
state space
optimal solution
loss function
matrix factorization
markov chain monte carlo
optical flow
parameter estimation
machine learning
prior knowledge
queueing networks