Matus Telgarsky
26 papers · 2010–2025 · 6 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+9 more ↓ Show less ↑
π§ Keyword Pioneer π Cross-Pollinator (10) π Conference Polyglot (6) π Academic Marathon (15) π Renaissance Researcher (5)
π
Renaissance Researcher
(5)
π
Interdisciplinary Bridge
πΊοΈ
Taxonomy Completionist
(21)
πΊ
Lone Wolf
(7)
π
Keyword Champion
(2)
π
Century Club
(26)
ποΈ
Keyword Collector
(68)
π₯
Unstoppable
(7)
π
Conference Pioneer
Conferences
COLT (8)
ICLR (7)
ICML (6)
AISTATS (2)
JMLR (2)
ALT (1)
Top co-authors
Keywords
gradient descent
(5)
empirical risk minimization
(3)
convex optimization
(3)
convergence rate
(3)
optimal transport
(2)
logistic regression
(2)
approximation theory
(2)
mirror descent
(2)
generalization bound
(2)
logistic loss
(2)
dual optimization
(2)
implicit bia
(2)
stochastic optimization
(2)
margin maximization
(2)
maximum margin
(2)
convergence analysis
(1)
primal-dual optimization
(1)
non-convex optimization
(1)
batch normalization
(1)
statistical consistency
(1)
Papers
Benefits of Early Stopping in Gradient Descent for Overparameterized Logistic Regression
ICML 2025
Spectrum Extraction and Clipping for Implicitly Linear Layers
AISTATS 2024
Large Stepsize Gradient Descent for Logistic Loss: Non-Monotonicity of the Loss Improves Optimization Efficiency
COLT 2024
Transformers, parallel computation, and logarithmic depth
ICML 2024
On Achieving Optimal Adversarial Test Error
ICLR 2023
Feature selection and low test error in shallow low-rotation ReLU networks
ICLR 2023
Actor-critic is implicitly biased towards high entropy optimal policies
ICLR 2022
Stochastic linear optimization never overfits with quadratically-bounded losses on general data
COLT 2022
Fast margin maximization via dual acceleration
ICML 2021
Characterizing the implicit bias via a primal-dual analysis
ALT 2021
Generalization bounds via distillation
ICLR 2021
Neural tangent kernels, transportation mappings, and universal approximation
ICLR 2020
Polylogarithmic width suffices for gradient descent to achieve arbitrarily small test error with shallow ReLU networks
ICLR 2020
Gradient descent follows the regularization path for general losses
COLT 2020
Gradient descent aligns the layers of deep linear networks
ICLR 2019
A Gradual, Semi-Discrete Approach to Generative Network Training via Explicit Wasserstein Minimization
ICML 2019
The implicit bias of gradient descent on nonseparable data
COLT 2019
Neural Networks and Rational Functions
ICML 2017
Non-convex learning via Stochastic Gradient Langevin Dynamics: a nonasymptotic analysis
COLT 2017
benefits of depth in neural networks
COLT 2016
Convex Risk Minimization and Conditional Probability Estimation
COLT 2015
Tensor Decompositions for Learning Latent Variable Models
JMLR 2014
Margins, Shrinkage, and Boosting
ICML 2013
Boosting with the Logistic Loss is Consistent
COLT 2013
A Primal-Dual Convergence Analysis of Boosting
JMLR 2012
Hartiganβs Method: k-means Clustering without Voronoi
AISTATS 2010