Variance-reduced First-order Methods for Finite-sum Function Constrained Convex Optimization
This paper studies finite-sum convex optimization with convex function constraints, where both the objective $f(x)=\frac{1}{m}\sum_{i=1}^m f_i(x)$ and the constraint mapping $g(x)=\frac{1}{n}\sum_{j=1}^n g_j(x)$ may have a large number of components. We propose a variance-reduced anchored constraint extrapolation method (VaRAn-ConEx), which combines stochastic component sampling with primal--dual extrapolation for nonlinear function constraints. Relative to stochastic approximation schemes such as ConEx, VaRAn-ConEx~method reduces gradient complexity by using epoch-wise full-gradient…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2611.00317
curl -X POST /api/v1/records/arxiv/2611.00317/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.00317/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'