no code implementations • 18 Dec 2023 • Alexander Bastounis, Felipe Cucker, Anders C. Hansen
However, we define a LASSO condition number and design an efficient algorithm for computing these support sets provided the input data is well-posed (has finite condition number) in time polynomial in the dimensions and logarithm of the condition number.
no code implementations • 13 Sep 2023 • Alexander Bastounis, Alexander N. Gorban, Anders C. Hansen, Desmond J. Higham, Danil Prokhorov, Oliver Sutton, Ivan Y. Tyukin, Qinghua Zhou
We consider classical distribution-agnostic framework and algorithms minimising empirical risks and potentially subjected to some weights regularisation.
no code implementations • 7 Sep 2023 • Oliver J. Sutton, Qinghua Zhou, Ivan Y. Tyukin, Alexander N. Gorban, Alexander Bastounis, Desmond J. Higham
We introduce a simple generic and generalisable framework for which key behaviours observed in practical systems arise with high probability -- notably the simultaneous susceptibility of the (otherwise accurate) model to easily constructed adversarial attacks, and robustness to random perturbations of the input data.
no code implementations • 13 Sep 2021 • Alexander Bastounis, Anders C Hansen, Verner Vlačić
Our paper addresses why there has been no solution to the problem, as we prove the following mathematical paradox: any training procedure based on training neural networks for classification problems with a fixed architecture will yield neural networks that are either inaccurate or unstable (if accurate) -- despite the provable existence of both accurate and stable neural networks for the same classification problems.
no code implementations • 26 Jun 2021 • Ivan Y. Tyukin, Desmond J. Higham, Alexander Bastounis, Eliyas Woldegeorgis, Alexander N. Gorban
Such a stealth attack could be conducted by a mischievous, corrupt or disgruntled member of a software development team.