Anastasis Kratsios

is this you? claim profile

∙ 0 followers
ETH Zurich ∙ Concordia University

PhD at Concordia University

Quantitative Rates and Fundamental Obstructions to Non-Euclidean Universal Approximation with Deep Narrow Feed-Forward Networks

By incorporating structured pairs of non-trainable input and output laye...
01/13/2021 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Optimizing Optimizers: Regret-optimal gradient descent algorithms

The need for fast and robust optimization algorithms are of critical imp...
12/31/2020 ∙ by Philippe Casgrain, et al. ∙ 0 ∙ share

read it
Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes

The effectiveness of neural networks in solving complex problems is well...
10/29/2020 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time

We introduce a simple neural network architecture modification that enab...
06/24/2020 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Non-Euclidean Universal Approximation

Modifications to a neural network's input and output layers are often re...
06/03/2020 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Universal Approximation Theorems

The universal approximation theorem established the density of specific ...
10/08/2019 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
The NEU Meta-Algorithm for Geometric Learning with Applications in Finance

We introduce a meta-algorithm, called non-Euclidean upgrading (NEU), whi...
08/31/2018 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Geometric Learning and Filtering in Finance

We develop a method for incorporating relevant non-Euclidean geometric i...
10/16/2017 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it
Arbitrage-Free Regularization

We introduce a path-dependent geometric framework which generalizes the ...
10/14/2017 ∙ by Anastasis Kratsios, et al. ∙ 0 ∙ share

read it

Anastasis Kratsios

Featured Co-authors

Philippe Casgrain

Cody B. Hyndman

Behnoosh Zamanlooy

Leonie Papon

Ievgen Bilokopytov

Quantitative Rates and Fundamental Obstructions to Non-Euclidean Universal Approximation with Deep Narrow Feed-Forward Networks

Optimizing Optimizers: Regret-optimal gradient descent algorithms

Overcoming The Limitations of Neural Networks in Composite-Pattern Learning with Architopes

Architopes: An Architecture Modification for Composite Pattern Learning, Increased Expressiveness, and Reduced Training Time

Non-Euclidean Universal Approximation

Universal Approximation Theorems

The NEU Meta-Algorithm for Geometric Learning with Applications in Finance

Geometric Learning and Filtering in Finance

Arbitrage-Free Regularization