@misc{indiciae51a8a0e0bdcd,
  title = {Langevin Monte-Carlo Provably Learns Depth Two Neural Nets at Any Size and Data},
  author = {Dibyakanti Kumar and Samyak Jha and Anirbit Mukherjee},
  year = {2025},
  url = {https://arxiv.org/abs/2503.10428},
  note = {Source identifier: 2503.10428}
}