Mikhail (Misha) Belkin
University of California, San Diego
H-index: 51
North America-United States
Top articles of Mikhail (Misha) Belkin
Title | Journal | Author(s) | Publication Date |
---|---|---|---|
Uncertainty Estimation with Recursive Feature Machines | Daniel Gedon Amirhesam Abedsoltan Thomas B Schön Mikhail Belkin | 2024 | |
Mechanism for feature learning in neural networks and backpropagation-free machine learning models | Science | Adityanarayanan Radhakrishnan Daniel Beaglehole Parthe Pandit Mikhail Belkin | 2024/3/7 |
The necessity of machine learning theory in mitigating AI risk. | ACM/JMS Journal of Data Science | Mikhail Belkin | 2024 |
Average gradient outer product as a mechanism for deep neural collapse | arXiv preprint arXiv:2402.13728 | Daniel Beaglehole Peter Súkeník Marco Mondelli Mikhail Belkin | 2024/2/21 |
Unmemorization in Large Language Models via Self-Distillation and Deliberate Imagination | arXiv preprint arXiv:2402.10052 | Yijiang River Dong Hongzhou Lin Mikhail Belkin Ramon Huerta Ivan Vulić | 2024/2/15 |
Aiming towards the minimizers: fast convergence of SGD for overparametrized problems | Advances in neural information processing systems | Chaoyue Liu Dmitriy Drusvyatskiy Misha Belkin Damek Davis Yian Ma | 2024/2/13 |
Linear Recursive Feature Machines provably recover low-rank matrices | arXiv preprint arXiv:2401.04553 | Adityanarayanan Radhakrishnan Mikhail Belkin Dmitriy Drusvyatskiy | 2024/1/9 |
On the Nyström Approximation for Preconditioning in Kernel Machines | Amirhesam Abedsoltan Parthe Pandit Luis Rademacher Mikhail Belkin | 2024/4/18 | |
Catapults in sgd: spikes in the training loss and their impact on generalization through feature learning | arXiv preprint arXiv:2306.04815 | Libin Zhu Chaoyue Liu Adityanarayanan Radhakrishnan Mikhail Belkin | 2023/6/7 |
More is Better: when Infinite Overparameterization is Optimal and Overfitting is Obligatory | James B Simon Dhruva Karkada Nikhil Ghosh Mikhail Belkin | 2023/10/13 | |
On emergence of clean-priority learning in early stopped neural networks | arXiv preprint arXiv:2306.02533 | Chaoyue Liu Amirhesam Abedsoltan Mikhail Belkin | 2023/6/5 |
Mechanism of clean-priority learning in early stopped neural networks of infinite width | Chaoyue Liu Amirhesam Abedsoltan Mikhail Belkin | 2023/10/13 | |
Wide and deep neural networks achieve consistency for classification | Proceedings of the National Academy of Sciences | Adityanarayanan Radhakrishnan Mikhail Belkin Caroline Uhler | 2023/4/4 |
A universal trade-off between the model size, test loss, and training loss of linear predictors | SIAM Journal on Mathematics of Data Science | Nikhil Ghosh Mikhail Belkin | 2023/12/31 |
SGD batch saturation for training wide neural networks | Chaoyue Liu Dmitriy Drusvyatskiy Mikhail Belkin Damek Davis Yian Ma | 2023/10/13 | |
Cut your losses with squentropy | Like Hui Mikhail Belkin Stephen Wright | 2023/7/3 | |
On the inconsistency of kernel ridgeless regression in fixed dimensions | SIAM Journal on Mathematics of Data Science | Daniel Beaglehole Mikhail Belkin Parthe Pandit | 2023/12/31 |
Mechanism of feature learning in convolutional neural networks | arXiv preprint arXiv:2309.00570 | Daniel Beaglehole Adityanarayanan Radhakrishnan Parthe Pandit Mikhail Belkin | 2023/9/1 |
On Feature Learning of Recursive Feature Machines and Automatic Relevance Determination | Daniel Gedon Amirhesam Abedsoltan Thomas B Schön Mikhail Belkin | 2023/12/18 | |
Toward large kernel models | ICML 2023 | Amirhesam Abedsoltan Mikhail Belkin Parthe Pandit | 2023/2/6 |