Acceleration by Stepsize Hedging: Multi-Step Descent and the Silver Stepsize Schedule
Can we accelerate the convergence of gradient descent without changing the algorithmÐjust by judiciously choosing stepsizes? Surprisingly, we show that the answer is yes. Our proposed Silver Stepsize Schedule optimizes strongly convex functions in � log� 2 ≈ � 0.7864 iterations, where � = 1 +...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
ACM
2025
|
| Online Access: | https://hdl.handle.net/1721.1/158132 |