AI RESEARCH
Convex Basins in Single-Index Model Loss Landscapes: Applications to Robust Recovery under Strong Adversarial Corruption
arXiv CS.LG
•
ArXi:2605.29497v1 Announce Type: new We study the problem of robustly learning Gaussian Single Index Models (SIMs) in the presence of heavy-tailed noise and a constant fraction of adversarially corrupted covariates and responses. Prior work on robust recovery has considered settings such as linear regression (Pensia, JASA 2024), strictly monotonic link functions (Awasthi, NeurIPS 2022), and phase retrieval (Buna and Rebeschini, AISTATS 2025