REVIEW 3 cited by
Tuning-Free Stochastic Optimization
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Large-scale machine learning problems make the cost of hyperparameter tuning ever more prohibitive. This creates a need for algorithms that can tune themselves on-the-fly. We formalize the notion of "tuning-free" algorithms that can match the performance of optimally-tuned optimization algorithms up to polylogarithmic factors given only loose hints on the relevant problem parameters. We consider in particular algorithms that can match optimally-tuned Stochastic Gradient Descent (SGD). When the domain of optimization is bounded, we show tuning-free matching of SGD is possible and achieved by several existing algorithms. We prove that for the task of minimizing a convex and smooth or Lipschitz function over an unbounded domain, tuning-free optimization is impossible. We discuss conditions under which tuning-free optimization is possible even over unbounded domains. In particular, we show that the recently proposed DoG and DoWG algorithms are tuning-free when the noise distribution is sufficiently well-behaved. For the task of finding a stationary point of a smooth and potentially nonconvex function, we give a variant of SGD that matches the best-known high-probability convergence rate for tuned SGD at only an additional polylogarithmic cost. However, we also give an impossibility result that shows no algorithm can hope to match the optimal expected convergence rate for tuned SGD with high probability.
Forward citations
Cited by 3 Pith papers
-
Randomized Feasibility Methods for Constrained Optimization with Adaptive Step Sizes
Combining adaptive DoWG-style step sizes with randomized Polyak feasibility updates yields a projection-free constrained optimization method with optimal O(1/√T) rates for convex objectives and linear convergence up t...
-
AutoSGD: Automatic Learning Rate Selection for Stochastic Gradient Descent
AutoSGD runs three parallel SGD streams at nearby learning rates, uses paired noisy objective estimates to pick the winner, and is claimed to converge with little user tuning.
-
A Parameter-Free and Near-Optimal Zeroth-Order Algorithm for Stochastic Convex Optimization
POEM is a parameter-free stochastic zeroth-order method that adapts both step size and smoothing automatically and reaches near-optimal oracle complexity.
Discussion (0). Continue with ORCID to comment.