Optimal Recursive Composition and Dyadic Phase Laws for Gradient Descent with Predetermined Stepsizes

Source: arxiv.org/abs/2609.11788

Yu Liu, Kang Chen, Rujun Jiang, Tianyu Wang

arXivarxiv/2609.11788Sep 10, 2026math.OCscore 0

Predetermined stepsize schedules featuring carefully chosen long steps have recently been shown to accelerate gradient descent (GD) on smooth convex functions. A prominent class of such schedules is built through recursive composition. In this paper, we characterize the convergence of these optimized recursive schedules, revealing a nonconstant log-periodic modulation across prescribed horizons. Specifically, we prove that for every $N \geq 1$, the optimized ConPP/OBS-S schedules satisfy $$f(x_{N-1})-f_\ast \le \frac{1}{2N^pΦ(\log_2N)-1} \frac{L}{2}\|x_0-x_\ast\|^2, \qquad…

JSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2609.11788
curl -X POST /api/v1/records/arxiv/2609.11788/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2609.11788/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'