Adaptive Stochastic Variance Reduction for Non-convex Finite-Sum Minimization
Kavis, Ali, Skoulakis, Stratis, Antonakopoulos, Kimon, Dadi, Leello Tadesse, Cevher, Volkan
–arXiv.org Artificial Intelligence
We propose an adaptive variance-reduction method, called AdaSpider, for minimization of $L$-smooth, non-convex functions with a finite-sum structure. In essence, AdaSpider combines an AdaGrad-inspired [Duchi et al., 2011, McMahan & Streeter, 2010], but a fairly distinct, adaptive step-size schedule with the recursive stochastic path integrated estimator proposed in [Fang et al., 2018]. To our knowledge, Adaspider is the first parameter-free non-convex variance-reduction method in the sense that it does not require the knowledge of problem-dependent parameters, such as smoothness constant $L$, target accuracy $\epsilon$ or any bound on gradient norms. In doing so, we are able to compute an $\epsilon$-stationary point with $\tilde{O}\left(n + \sqrt{n}/\epsilon^2\right)$ oracle-calls, which matches the respective lower bound up to logarithmic factors.
arXiv.org Artificial Intelligence
Nov-3-2022
- Country:
- North America
- United States
- New York
- Richmond County > New York City (0.04)
- Queens County > New York City (0.04)
- New York County > New York City (0.04)
- Kings County > New York City (0.04)
- Bronx County > New York City (0.04)
- Colorado > Boulder County
- Boulder (0.04)
- California > Los Angeles County
- Long Beach (0.04)
- New York
- Puerto Rico > San Juan
- San Juan (0.04)
- Canada
- Quebec > Montreal (0.04)
- British Columbia > Metro Vancouver Regional District
- Vancouver (0.04)
- United States
- Europe
- Austria (0.04)
- United Kingdom > England
- Greater London > London (0.04)
- Switzerland > Vaud
- Lausanne (0.04)
- Spain > Catalonia
- Barcelona Province > Barcelona (0.04)
- Asia
- Middle East > Israel
- Haifa District > Haifa (0.04)
- Japan > Kyūshū & Okinawa
- Okinawa (0.04)
- Middle East > Israel
- Africa > Ethiopia
- Addis Ababa > Addis Ababa (0.04)
- North America
- Genre:
- Research Report (1.00)
- Technology: