Acceleration Survives Heavy Tails and Generalized Smoothness
We establish high-probability accelerated convergence guarantees for convex $(L_0,L_1)$-smooth optimization with unbiased stochastic gradients under heavy-tailed noise, assuming a central $α$-th noise moment bounded by $σ^α$, $α\in(1,2]$. Such guarantees are known under classical $L$-smoothness, but extending them through localization over an $O(R)$-radius ball may introduce an $e^{O(L_1R)}$ factor. We avoid this dependence in two ways. First, a two-phase scheme uses Clip-SGD to reach a low-curvature regime, after which clipped stochastic similar triangles (Clip-SSTM) attains, with…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2610.23301
curl -X POST /api/v1/records/arxiv/2610.23301/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2610.23301/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'