MS student, Tel Aviv University
1 paper at NeurIPS 2025
We show that clipped SGD converges with high probability on convex $(L_0,L_1)$-smooth functions.