Mikhail Belkin

Publication Activity (10 Years)

Years Active: 2001-2024
Publications (10 Years): 87

Top Venues

CoRR

NIPS

ICML

ICLR

Publications

James B. Simon, Dhruva Karkada, Nikhil Ghosh, Mikhail Belkin
More is Better: when Infinite Overparameterization is Optimal and Overfitting is Obligatory. ICLR (2024)
Libin Zhu, Chaoyue Liu, Adityanarayanan Radhakrishnan, Mikhail Belkin
Quadratic models for understanding catapult dynamics of neural networks. ICLR (2024)
Neil Mallinar, Daniel Beaglehole, Libin Zhu, Adityanarayanan Radhakrishnan, Parthe Pandit, Mikhail Belkin
Emergence in non-neural models: grokking modular arithmetic via average gradient outer product. CoRR (2024)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Dmitriy Drusvyatskiy
Linear Recursive Feature Machines provably recover low-rank matrices. CoRR (2024)
Amirhesam Abedsoltan, Parthe Pandit, Luis Rademacher, Mikhail Belkin
On the Nyström Approximation for Preconditioning in Kernel Machines. AISTATS (2024)
Daniel Beaglehole, Peter Súkeník, Marco Mondelli, Mikhail Belkin
Average gradient outer product as a mechanism for deep neural collapse. CoRR (2024)
Yijiang River Dong, Hongzhou Lin, Mikhail Belkin, Ramón Huerta, Ivan Vulic
Unmemorization in Large Language Models via Self-Distillation and Deliberate Imagination. CoRR (2024)
James B. Simon, Dhruva Karkada, Nikhil Ghosh, Mikhail Belkin
More is Better in Modern Machine Learning: when Infinite Overparameterization is Optimal and Overfitting is Obligatory. CoRR (2023)
Like Hui, Mikhail Belkin, Stephen Wright
Cut your Losses with Squentropy. CoRR (2023)
Amirhesam Abedsoltan, Mikhail Belkin, Parthe Pandit
Toward Large Kernel Models. ICML (2023)
Chaoyue Liu, Amirhesam Abedsoltan, Mikhail Belkin
On Emergence of Clean-Priority Learning in Early Stopped Neural Networks. CoRR (2023)
Daniel Beaglehole, Adityanarayanan Radhakrishnan, Parthe Pandit, Mikhail Belkin
Mechanism of feature learning in convolutional neural networks. CoRR (2023)
Amirhesam Abedsoltan, Mikhail Belkin, Parthe Pandit, Luis Rademacher
On the Nystrom Approximation for Preconditioning in Kernel Machines. CoRR (2023)
Amirhesam Abedsoltan, Mikhail Belkin, Parthe Pandit
Toward Large Kernel Models. CoRR (2023)
Nikhil Ghosh, Mikhail Belkin
A Universal Trade-off Between the Model Size, Test Loss, and Training Loss of Linear Predictors. SIAM J. Math. Data Sci. 5 (4) (2023)
Arindam Banerjee, Pedro Cisneros-Velarde, Libin Zhu, Mikhail Belkin
Neural tangent kernel at initialization: linear width suffices. UAI (2023)
Chaoyue Liu, Dmitriy Drusvyatskiy, Mikhail Belkin, Damek Davis, Yi-An Ma
Aiming towards the minimizers: fast convergence of SGD for overparametrized problems. CoRR (2023)
Libin Zhu, Chaoyue Liu, Adityanarayanan Radhakrishnan, Mikhail Belkin
Catapults in SGD: spikes in the training loss and their impact on generalization through feature learning. CoRR (2023)
Daniel Beaglehole, Mikhail Belkin, Parthe Pandit
On the Inconsistency of Kernel Ridgeless Regression in Fixed Dimensions. SIAM J. Math. Data Sci. 5 (4) (2023)
Like Hui, Mikhail Belkin, Stephen Wright
Cut your Losses with Squentropy. ICML (2023)
Neil Mallinar, James B. Simon, Amirhesam Abedsoltan, Parthe Pandit, Mikhail Belkin, Preetum Nakkiran
Benign, Tempered, or Catastrophic: A Taxonomy of Overfitting. CoRR (2022)
Libin Zhu, Parthe Pandit, Mikhail Belkin
A note on Linear Bottleneck networks and their Transition to Multilinearity. CoRR (2022)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Wide and Deep Neural Networks Achieve Optimality for Classification. CoRR (2022)
Adityanarayanan Radhakrishnan, Daniel Beaglehole, Parthe Pandit, Mikhail Belkin
Feature learning in neural networks and kernel machines that recursively learn features. CoRR (2022)
Yuan Cao, Zixiang Chen, Mikhail Belkin, Quanquan Gu
Benign Overfitting in Two-layer Convolutional Neural Networks. CoRR (2022)
Like Hui, Mikhail Belkin, Preetum Nakkiran
Limitations of Neural Collapse for Understanding Generalization in Deep Learning. CoRR (2022)
Libin Zhu, Chaoyue Liu, Mikhail Belkin
Transition to Linearity of General Neural Networks with Directed Acyclic Graph Architecture. CoRR (2022)
Chaoyue Liu, Libin Zhu, Mikhail Belkin
Transition to Linearity of Wide Neural Networks is an Emerging Property of Assembling Weak Models. CoRR (2022)
Arindam Banerjee, Pedro Cisneros-Velarde, Libin Zhu, Mikhail Belkin
Restricted Strong Convexity of Deep Learning Models with Smooth Activations. CoRR (2022)
Nikhil Ghosh, Mikhail Belkin
A Universal Trade-off Between the Model Size, Test Loss, and Training Loss of Linear Predictors. CoRR (2022)
Daniel Beaglehole, Mikhail Belkin, Parthe Pandit
Kernel Ridgeless Regression is Inconsistent for Low Dimensions. CoRR (2022)
Libin Zhu, Chaoyue Liu, Adityanarayanan Radhakrishnan, Mikhail Belkin
Quadratic models for understanding neural network dynamics. CoRR (2022)
Adityanarayanan Radhakrishnan, George Stefanakis, Mikhail Belkin, Caroline Uhler
Simple, Fast, and Flexible Framework for Matrix Completion with Infinite Width Neural Networks. CoRR (2021)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Local Quadratic Convergence of Stochastic Gradient Descent with Adaptive Step Size. CoRR (2021)
Lin Chen, Yifei Min, Mikhail Belkin, Amin Karbasi
Multiple Descent: Design Your Own Generalization Curve. NeurIPS (2021)
Yuan Cao, Quanquan Gu, Mikhail Belkin
Risk Bounds for Over-parameterized Maximum Margin Classification on Sub-Gaussian Mixtures. CoRR (2021)
Mikhail Belkin
Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolation. CoRR (2021)
Mikhail Belkin
Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolation. Acta Numer. 30 (2021)
Yuan Cao, Quanquan Gu, Mikhail Belkin
Risk Bounds for Over-parameterized Maximum Margin Classification on Sub-Gaussian Mixtures. NeurIPS (2021)
Vidya Muthukumar, Adhyyan Narang, Vignesh Subramanian, Mikhail Belkin, Daniel Hsu, Anant Sahai
Classification vs regression in overparameterized regimes: Does the loss function matter? J. Mach. Learn. Res. 22 (2021)
Like Hui, Mikhail Belkin
Evaluation of Neural Architectures trained with square Loss vs Cross-Entropy in Classification Tasks. ICLR (2021)
Qichao Que, Mikhail Belkin
Back to the Future: Radial Basis Function Network Revisited. IEEE Trans. Pattern Anal. Mach. Intell. 42 (8) (2020)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Linear Convergence and Implicit Regularization of Generalized Mirror Descent with Time-Dependent Mirrors. CoRR (2020)
Vidya Muthukumar, Adhyyan Narang, Vignesh Subramanian, Mikhail Belkin, Daniel J. Hsu, Anant Sahai
Classification vs regression in overparameterized regimes: Does the loss function matter? CoRR (2020)
Chaoyue Liu, Libin Zhu, Mikhail Belkin
On the linearity of large non-linear models: when and why the tangent kernel is constant. CoRR (2020)
Chaoyue Liu, Libin Zhu, Mikhail Belkin
Toward a theory of optimization for over-parameterized systems of non-linear equations: the lessons of deep learning. CoRR (2020)
Lin Chen, Yifei Min, Mikhail Belkin, Amin Karbasi
Multiple Descent: Design Your Own Generalization Curve. CoRR (2020)
Like Hui, Mikhail Belkin
Evaluation of Neural Architectures Trained with Square Loss vs Cross-Entropy in Classification Tasks. CoRR (2020)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Overparameterized neural networks implement associative memory. Proc. Natl. Acad. Sci. USA 117 (44) (2020)
Mikhail Belkin, Daniel Hsu, Ji Xu
Two Models of Double Descent for Weak Features. SIAM J. Math. Data Sci. 2 (4) (2020)
Chaoyue Liu, Libin Zhu, Mikhail Belkin
On the linearity of large non-linear models: when and why the tangent kernel is constant. NeurIPS (2020)
Chaoyue Liu, Mikhail Belkin
Accelerating SGD with momentum for over-parameterized learning. ICLR (2020)
Mikhail Belkin, Daniel Hsu, Ji Xu
Two models of double descent for weak features. CoRR (2019)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Overparameterized Neural Networks Can Implement Associative Memory. CoRR (2019)
Like Hui, Siyuan Ma, Mikhail Belkin
Kernel Machines Beat Deep Neural Networks on Mask-Based Single-Channel Speech Enhancement. INTERSPEECH (2019)
Siyuan Ma, Mikhail Belkin
Kernel Machines That Adapt To Gpus For Effective Large Batch Training. MLSys (2019)
Mikhail Belkin, Alexander Rakhlin, Alexandre B. Tsybakov
Does data interpolation contradict statistical optimality? AISTATS (2019)
Mikhail Belkin, Alexander Rakhlin, Alexandre B. Tsybakov
Does data interpolation contradict statistical optimality? CoRR (2018)
Chaoyue Liu, Mikhail Belkin
Parametrized Accelerated Methods Free of Condition Number. CoRR (2018)
Siyuan Ma, Raef Bassily, Mikhail Belkin
The Power of Interpolation: Understanding the Effectiveness of SGD in Modern Over-parametrized Learning. ICML (2018)
Raef Bassily, Mikhail Belkin, Siyuan Ma
On exponential convergence of SGD in non-convex over-parametrized learning. CoRR (2018)
Mikhail Belkin, Daniel Hsu, Siyuan Ma, Soumik Mandal
Reconciling modern machine learning and the bias-variance trade-off. CoRR (2018)
Siyuan Ma, Mikhail Belkin
Learning kernels that adapt to GPU. CoRR (2018)
Mikhail Belkin
Approximation beats concentration? An approximation view on inference with smooth radial kernels. CoRR (2018)
Akshay Mehra, Jihun Hamm, Mikhail Belkin
Fast Interactive Image Retrieval using large-scale unlabeled data. CoRR (2018)
Mikhail Belkin, Siyuan Ma, Soumik Mandal
To understand deep learning we need to understand kernel learning. CoRR (2018)
Mikhail Belkin, Luis Rademacher, James R. Voss
Eigenvectors of Orthogonally Decomposable Functions. SIAM J. Comput. 47 (2) (2018)
Like Hui, Siyuan Ma, Mikhail Belkin
Kernel Machines Beat Deep Neural Networks on Mask-based Single-channel Speech Enhancement. CoRR (2018)
Mikhail Belkin
Approximation beats concentration? An approximation view on inference with smooth radial kernels. COLT (2018)
Mikhail Belkin, Daniel Hsu, Partha Mitra
Overfitting or perfect fitting? Risk bounds for classification and regression rules that interpolate. CoRR (2018)
Justin Eldridge, Mikhail Belkin, Yusu Wang
Unperturbed: spectral analysis beyond Davis-Kahan. ALT (2018)
Chaoyue Liu, Mikhail Belkin
MaSS: an Accelerated Stochastic Method for Over-parametrized Learning. CoRR (2018)
Mikhail Belkin, Siyuan Ma, Soumik Mandal
To Understand Deep Learning We Need to Understand Kernel Learning. ICML (2018)
Adityanarayanan Radhakrishnan, Mikhail Belkin, Caroline Uhler
Downsampling leads to Image Memorization in Convolutional Autoencoders. CoRR (2018)
Mikhail Belkin, Daniel J. Hsu, Partha Mitra
Overfitting or perfect fitting? Risk bounds for classification and regression rules that interpolate. NeurIPS (2018)
Siyuan Ma, Raef Bassily, Mikhail Belkin
The Power of Interpolation: Understanding the Effectiveness of SGD in Modern Over-parametrized Learning. CoRR (2017)
Siyuan Ma, Mikhail Belkin
Diving into the shallows: a computational perspective on large-scale shallow learning. CoRR (2017)
Siyuan Ma, Mikhail Belkin
Diving into the shallows: a computational perspective on large-scale shallow learning. NIPS (2017)
Justin Eldridge, Mikhail Belkin, Yusu Wang
Unperturbed: spectral analysis beyond Davis-Kahan. CoRR (2017)
Chaoyue Liu, Mikhail Belkin
Clustering with Bregman Divergences: an Asymptotic Analysis. NIPS (2016)
Mikhail Belkin, Luis Rademacher, James R. Voss
Basis Learning as an Algorithmic Primitive. COLT (2016)
Justin Eldridge, Mikhail Belkin, Yusu Wang
Graphons, mergeons, and so on! NIPS (2016)
Jihun Hamm, Paul Cao, Mikhail Belkin
Learning Privately from Multiparty Data. CoRR (2016)
Justin Eldridge, Mikhail Belkin, Yusu Wang
Graphons, mergeons, and so on! CoRR (2016)
Qichao Que, Mikhail Belkin
Back to the Future: Radial Basis Function Networks Revisited. AISTATS (2016)
James R. Voss, Mikhail Belkin, Luis Rademacher
The Hidden Convexity of Spectral Clustering. AAAI (2016)
Jihun Hamm, Yingjun Cao, Mikhail Belkin
Learning privately from multiparty data. ICML (2016)
Justin Eldridge, Mikhail Belkin, Yusu Wang
Beyond Hartigan Consistency: Merge Distortion Metric for Hierarchical Clustering. COLT (2015)
James R. Voss, Mikhail Belkin, Luis Rademacher
Optimal Recovery in Noisy ICA. CoRR (2015)
Mikhail Belkin, Vladimir Iakovlev
Microwave-Band Circuit-Level Semiconductor Laser Modeling. EMS (2015)
Jihun Hamm, Mikhail Belkin
Probabilistic Zero-shot Classification with Semantic Rankings. CoRR (2015)
Jihun Hamm, Adam C. Champion, Guoxing Chen, Mikhail Belkin, Dong Xuan
Crowd-ML: A Privacy-Preserving Learning Framework for a Crowd of Smart Devices. CoRR (2015)
Jihun Hamm, Adam C. Champion, Guoxing Chen, Mikhail Belkin, Dong Xuan
Crowd-ML: A Privacy-Preserving Learning Framework for a Crowd of Smart Devices. ICDCS (2015)
James R. Voss, Mikhail Belkin, Luis Rademacher
A Pseudo-Euclidean Iteration for Optimal Recovery in Noisy ICA. NIPS (2015)
Mikhail Belkin, Kaushik Sinha
Polynomial Learning of Distribution Families. SIAM J. Comput. 44 (4) (2015)
Qichao Que, Mikhail Belkin, Yusu Wang
Learning with Fredholm Kernels. NIPS (2014)
Mikhail Belkin, Luis Rademacher, James R. Voss
The Hidden Convexity of Spectral Clustering. CoRR (2014)
Joseph Anderson, Mikhail Belkin, Navin Goyal, Luis Rademacher, James R. Voss
The More, the Merrier: the Blessing of Dimensionality for Learning Large Gaussian Mixtures. COLT (2014)
Mikhail Belkin, Luis Rademacher, James R. Voss
Learning a Hidden Basis Through Imperfect Measurements: An Algorithmic Primitive. CoRR (2014)
James R. Voss, Luis Rademacher, Mikhail Belkin
Fast Algorithms for Gaussian Noise Invariant Independent Component Analysis. NIPS (2013)

Mikhail Belkin.css-eqlu8i{font-size:small;padding-left:4px;}

Publication Activity (10 Years)

Top Topics

Top Venues

Mikhail Belkin