Följ
Alireza F. Pour
Alireza F. Pour
University of Waterloo
Verifierad e-postadress på uwaterloo.ca
Titel
Citeras av
Citeras av
År
Benefits of Additive Noise in Composing Classes with Bounded Capacity
AF Pour, H Ashtiani
Advances in Neural Information Processing Systems 35, 32709-32722, 2022
32022
Sample-Optimal Locally Private Hypothesis Selection and the Provable Benefits of Interactivity
AF Pour, H Ashtiani, S Asoodeh
arXiv preprint arXiv:2312.05645, 2023
2023
On the Role of Noise in the Sample Complexity of Learning Recurrent Neural Networks: Exponential Gaps for Long Sequences
AF Pour, H Ashtiani
Advances in Neural Information Processing Systems 36, 2023
2023
Benefits of Additive Noise in Composing Classes of Functions with Applications to Neural Networks
A Fathollah Pour
2022
Systemet kan inte utföra åtgärden just nu. Försök igen senare.
Artiklar 1–4