Optimal Last-Iterate Bounds for Nonexpansive Stochastic Approximation
Source: arxiv.org/abs/2611.13846
We establish last-iterate mean-square residual bounds for nonexpansive stochastic approximation (SA), or stochastic Krasnoselskii-Mann (KM) iterations. In particular, we consider fixed-point mappings that are nonexpansive with respect to the Euclidean norm, and focus on polynomially decaying stepsizes. We study three conditions on the conditional second moment of the noise, and obtain bounds that are optimal up to logarithmic factors in each case. When it scales linearly with the squared residual, the best rate is $O(1/k^{1-ε})$ for any $ε>0$. In this case, for all considered stepsizes, our…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2611.13846
curl -X POST /api/v1/records/arxiv/2611.13846/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.13846/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'