Amit Saha

Current Work

I'm currently focusing on the following projects:
  1. Privacy amplificiation via composite subsampling schemes for learning from structured data
  2. Efficient sampling mechanisms for private optimizers (e.g. Muon) in high dimension
  3. Memorization and generalization in diffusion models
  4. Frameworks + provable guarantees for algorithmic copyright protection in generative modelling

Conference Papers

  1. Chien, E., Saha, A., Huang, Y., & Li, P. "GenAI Copyright Evidence with Operational Meaning", International Conference on Machine Learning (ICML), MemFM + R2FM Workshops , June 2025.
  2. Govil, S., Rodgers, J. P., Chou, Y.-T., Miao, S., Saha, A., Anand, A., Lieret, K., DeZoort, G., Liu, M., Duarte, J., Li, P., & Hsu, S.-C. "Locality-Sensitive Hashing-Based Efficient Point Transformer for Charged Particle Reconstruction", NeurIPS 2025 Workshop on Machine Learning and the Physical Sciences, Dec 2025.