Efficient Convex Optimization Requires Superlinear Memory
Marsden, Annie, Sharan, Vatsal, Sidford, Aaron, Valiant, Gregory
We show that any memory-constrained, first-order algorithm which minimizes $d$-dimensional, $1$-Lipschitz convex functions over the unit ball to $1/\mathrm{poly}(d)$ accuracy using at most $d^{1.25 - \delta}$ bits of memory must make at least $\tilde{\Omega}(d^{1 + (4/3)\delta})$ first-order queries (for any constant $\delta \in [0, 1/4]$). Consequently, the performance of such memory-constrained algorithms are a polynomial factor worse than the optimal $\tilde{O}(d)$ query bound for this problem obtained by cutting plane methods that use $\tilde{O}(d^2)$ memory. This resolves a COLT 2019 open problem of Woodworth and Srebro.
Mar-29-2022
- Country:
- North America > United States
- Virginia (0.04)
- California > Santa Clara County
- Palo Alto (0.04)
- Europe
- Russia (0.04)
- United Kingdom > England
- Cambridgeshire > Cambridge (0.04)
- Asia
- Middle East > Jordan (0.04)
- Russia (0.04)
- North America > United States
- Genre:
- Research Report (0.50)
- Technology: