Abstract: Modern neural network training relies on piece-wise (sub-)differentiable functions in order to use backpropagation to update model parameters. In this work, we introduce a novel method to ...

This repository contains a full solution for training a machine learning model using a non-differentiable loss function. It satisfies the academic requirements by: ...

We show how complexity theory can be introduced in machine learning to help bring together apparently disparate areas of current research. We show that this model-driven approach may require less ...

This paper aims to present a multifractal approach of the turbulent atmosphere, by proposing that it can be considered a complex system whose structural units support dynamics on continuous but ...

In this paper we bound character sums of the shape ∑ n≤N χ 1 f n χ 2 f n+l , ; where χ1 and χ2 are non-principal multiplicative characters modulo a prime p, f(x) is a real-valued, twice-differentiable ...