Announcement_24_sparsetraining_icml

Adnan Mohammed and Rohan Jain’s work on “Sparse Training from Random Initialization: Aligning Lottery Ticket Masks using Weight Symmetry” (Adnan et al., 2025) has been accepted at the International Conference on Machine Learning (ICML), 2025. This work explores the Lottery Ticket Hypothesis (LTH) and sparse training from a random initialization through the lens of weight and permutation symmetry, proposing a novel approach to improve LTH mask generalization across new random initialization.