Anastasis Kratsios

Publication Activity (10 Years)

Years Active: 2018-2024
Publications (10 Years): 43

Top Topics

Kernel Ridge Regression

Generalization Bounds

Top Venues

J. Mach. Learn. Res.

Trans. Mach. Learn. Res.

Publications

Raeid Saqur, Anastasis Kratsios, Florian Krach, Yannick Limmer, Jacob-Junqi Tian, John Willes, Blanka Horvath, Frank Rudzicz
Filtered not Mixed: Stochastic Filtering-Based Online Gating for Mixture of Large Language Models. CoRR (2024)
Haitz Sáez de Ocáriz Borde, Anastasis Kratsios
Neural Snowflakes: Universal Latent Graph Inference via Trainable Latent Geometries. ICLR (2024)
Haitz Sáez de Ocáriz Borde, Takashi Furuya, Anastasis Kratsios, Marc T. Law
Breaking the Curse of Dimensionality with Distributed Neural Computation. CoRR (2024)
Tin Sum Cheng, Aurélien Lucchi, Anastasis Kratsios, David Belius
Characterizing Overfitting in Kernel Ridgeless Regression Through the Eigenspectrum. CoRR (2024)
Reza Arabpour, John Armstrong, Luca Galimberti, Anastasis Kratsios, Giulia Livieri
Low-dimensional approximations of the conditional law of Volterra processes: a non-positive curvature approach. CoRR (2024)
Anastasis Kratsios, Takashi Furuya, Jose Antonio Lara Benitez, Matti Lassas, Maarten V. de Hoop
Mixture of Experts Soften the Curse of Dimensionality in Operator Learning. CoRR (2024)
Anastasis Kratsios, A. Martina Neuman, Gudmund Pammer
Digital Computers Break the Curse of Dimensionality: Adaptive Bounds via Finite Geometry. CoRR (2024)
Jose Antonio Lara Benitez, Takashi Furuya, Florian Faucher, Anastasis Kratsios, Xavier Tricoche, Maarten V. de Hoop
Out-of-distributional risk bounds for neural operators with applications to the Helmholtz equation. J. Comput. Phys. 513 (2024)
Anastasis Kratsios, Ruiyang Hong, Haitz Sáez de Ocáriz Borde
Capacity bounds for hyperbolic neural network representations of latent tree structures. Neural Networks 178 (2024)
Yannick Limmer, Anastasis Kratsios, Xuwei Yang, Raeid Saqur, Blanka Horvath
Reality Only Happens Once: Single-Path Generalization Bounds for Transformers. CoRR (2024)
Anastasis Kratsios, Chong Liu, Matti Lassas, Maarten V. de Hoop, Ivan Dokmanic
A Transfer Principle: Universal Approximators Between Metric Spaces From Euclidean Universal Approximators. CoRR (2023)
Anastasis Kratsios, Cody B. Hyndman
Generative Ornstein-Uhlenbeck Markets via Geometric Deep Learning. CoRR (2023)
Anastasis Kratsios, Valentin Debarnot, Ivan Dokmanic
Small Transformers Compute Universal Metric Embeddings. J. Mach. Learn. Res. 24 (2023)
Calypso Herrera, Florian Krach, Anastasis Kratsios, Pierre Ruyssen, Josef Teichmann
Denise: Deep Robust Principal Component Analysis for Positive Semidefinite Matrices. Trans. Mach. Learn. Res. 2023 (2023)
Haitz Sáez de Ocáriz Borde, Anastasis Kratsios
Neural Snowflakes: Universal Latent Graph Inference via Trainable Latent Geometries. CoRR (2023)
Tin Sum Cheng, Aurélien Lucchi, Ivan Dokmanic, Anastasis Kratsios, David Belius
A Theoretical Analysis of the Test Error of Finite-Rank Kernel Ridge Regression. CoRR (2023)
Xuwei Yang, Anastasis Kratsios, Florian Krach, Matheus R. Grasselli, Aurélien Lucchi
Regret-Optimal Federated Transfer Learning for Kernel Regression with Applications in American Option Pricing. CoRR (2023)
Anastasis Kratsios, Ruiyang Hong, Haitz Sáez de Ocáriz Borde
Capacity Bounds for Hyperbolic Neural Network Representations of Latent Tree Structures. CoRR (2023)
Blanka Hovart, Anastasis Kratsios, Yannick Limmer, Xuwei Yang
Deep Kalman Filters Can Filter. CoRR (2023)
Tin Sum Cheng, Aurélien Lucchi, Anastasis Kratsios, Ivan Dokmanic, David Belius
A Theoretical Analysis of the Test Error of Finite-Rank Kernel Ridge Regression. NeurIPS (2023)
Anastasis Kratsios, Behnoosh Zamanlooy
Do ReLU Networks Have An Edge When Approximating Compactly-Supported Functions? Trans. Mach. Learn. Res. 2022 (2022)
Anastasis Kratsios, Behnoosh Zamanlooy
Learning sub-patterns in piecewise continuous functions. Neurocomputing 480 (2022)
Anastasis Kratsios, Behnoosh Zamanlooy
Piecewise-Linear Activations or Analytic Activation Functions: Which Produce More Expressive Neural Networks? CoRR (2022)
Anastasis Kratsios, Behnoosh Zamanlooy, Tianlin Liu, Ivan Dokmanic
Universal Approximation Under Constraints is Possible with Transformers. ICLR (2022)
Anastasis Kratsios, Leonie Papon
Universal Approximation Theorems for Differentiable Geometric Deep Learning. J. Mach. Learn. Res. 23 (2022)
Luca Galimberti, Giulia Livieri, Anastasis Kratsios
Designing Universal Causal Deep Learning Models: The Case of Infinite-Dimensional Dynamical Systems from Stochastic Analysis. CoRR (2022)
Beatrice Acciaio, Anastasis Kratsios, Gudmund Pammer
Metric Hypertransformers are Universal Adapted Maps. CoRR (2022)
Songyan Hou, Parnian Kassraie, Anastasis Kratsios, Jonas Rothfuss, Andreas Krause
Instance-Dependent Generalization Bounds via Optimal Transport. CoRR (2022)
Anastasis Kratsios, Valentin Debarnot, Ivan Dokmanic
Small Transformers Compute Universal Metric Embeddings. CoRR (2022)
Philippe Casgrain, Anastasis Kratsios
Optimizing Optimizers: Regret-optimal gradient descent algorithms. COLT (2021)
Philippe Casgrain, Anastasis Kratsios
Optimizing Optimizers: Regret-optimal gradient descent algorithms. CoRR (2021)
Anastasis Kratsios
Universal Regular Conditional Distributions via Probability Measure-Valued Deep Neural Models. CoRR (2021)
Anastasis Kratsios, Behnoosh Zamanlooy, Tianlin Liu, Ivan Dokmanic
Universal Approximation Under Constraints is Possible with Transformers. CoRR (2021)
Anastasis Kratsios, Leonie Papon
Quantitative Rates and Fundamental Obstructions to Non-Euclidean Universal Approximation with Deep Narrow Feed-Forward Networks. CoRR (2021)
Anastasis Kratsios, Cody B. Hyndman
Generative OrnsteinUhlenbeck Markets via Geometric Deep Learning. GSI (2) (2021)
Anastasis Kratsios, Cody B. Hyndman
NEU: A Meta-Algorithm for Universal UAP-Invariant Feature Representation. J. Mach. Learn. Res. 22 (2021)
Anastasis Kratsios
The Universal Approximation Property. Ann. Math. Artif. Intell. 89 (5-6) (2021)
Anastasis Kratsios, Ievgen Bilokopytov
Non-Euclidean Universal Approximation. NeurIPS (2020)
Anastasis Kratsios, Behnoosh Zamanlooy
Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time. CoRR (2020)
Anastasis Kratsios, Ievgen Bilokopytov
Non-Euclidean Universal Approximation. CoRR (2020)
Anastasis Kratsios, Behnoosh Zamanlooy
Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes. CoRR (2020)
Anastasis Kratsios
Universal Approximation Theorems. CoRR (2019)
Anastasis Kratsios, Cody B. Hyndman
The NEU Meta-Algorithm for Geometric Learning with Applications in Finance. CoRR (2018)