Showing 1–5 of 5 results for author: Bondarenko, I

The word problem and growth of groups

Authors: Ievgen Bondarenko

Abstract: Let $\mathrm{WP}_G$ denote the word problem in a finitely generated group $G$. We consider the complexity of $\mathrm{WP}_G$ with respect to standard deterministic Turing machines. Let $\mathrm{DTIME}_k(t(n))$ be the complexity class of languages solved in time $O(t(n))$ by a Turing machine with $k$ tapes. We prove that $\mathrm{WP}_G\in\mathrm{DTIME}_1(n\log n)$ if and only if $G$ is virtually ni… ▽ More Let $\mathrm{WP}_G$ denote the word problem in a finitely generated group $G$. We consider the complexity of $\mathrm{WP}_G$ with respect to standard deterministic Turing machines. Let $\mathrm{DTIME}_k(t(n))$ be the complexity class of languages solved in time $O(t(n))$ by a Turing machine with $k$ tapes. We prove that $\mathrm{WP}_G\in\mathrm{DTIME}_1(n\log n)$ if and only if $G$ is virtually nilpotent. We relate the complexity of the word problem and the growth of groups by showing that $\mathrm{WP}_G\not\in \mathrm{DTIME}_1(o(n\logγ(n)))$, where $γ(n)$ is the growth function of $G$. We prove that $\mathrm{WP}_G\in\mathrm{DTIME}_k(n)$ for strongly contracting automaton groups, $\mathrm{WP}_G\in\mathrm{DTIME}_k(n\log n)$ for groups generated by bounded automata, and $\mathrm{WP}_G\in\mathrm{DTIME}_k(n(\log n)^d)$ for groups generated by polynomial automata. In particular, for the Grigorchuk group, $\mathrm{WP}_G\not\in\mathrm{DTIME}_1(n^{1.7674})$ and $\mathrm{WP}_G\in\mathrm{DTIME}_1(n^2)$. △ Less

Submitted 17 March, 2024; originally announced March 2024.

Comments: 14 pages

MSC Class: 20F10; 68Q70; 03D10; 20E08

More layers! End-to-end regression and uncertainty on tabular data with deep learning

Authors: Ivan Bondarenko

Abstract: This paper attempts to analyze the effectiveness of deep learning for tabular data processing. It is believed that decision trees and their ensembles is the leading method in this domain, and deep neural networks must be content with computer vision and so on. But the deep neural network is a framework for building gradient-based hierarchical representations, and this key feature should be able to… ▽ More This paper attempts to analyze the effectiveness of deep learning for tabular data processing. It is believed that decision trees and their ensembles is the leading method in this domain, and deep neural networks must be content with computer vision and so on. But the deep neural network is a framework for building gradient-based hierarchical representations, and this key feature should be able to provide the best processing of generic structured (tabular) data, not just image matrices and audio spectrograms. This problem is considered through the prism of the Weather Prediction track in the Yandex Shifts challenge (in other words, the Yandex Shifts Weather task). This task is a variant of the classical tabular data regression problem. It is also connected with another important problem: generalization and uncertainty in machine learning. This paper proposes an end-to-end algorithm for solving the problem of regression with uncertainty on tabular data, which is based on the combination of four ideas: 1) deep ensemble of self-normalizing neural networks, 2) regression as parameter estimation of the Gaussian target error distribution, 3) hierarchical multitask learning, and 4) simple data preprocessing. Three modifications of the proposed algorithm form the top-3 leaderboard of the Yandex Shifts Weather challenge respectively. This paper considers that this success has occurred due to the fundamental properties of the deep learning algorithm, and tries to prove this. △ Less

Submitted 7 December, 2021; originally announced December 2021.

Comments: 12 pages, 5 figures, the described solution is submitted to the Shifts Challenge (see https://research.yandex.com/shifts/weather), the code is available on https://github.com/bond005/yandex-shifts-weather

MSC Class: 68T07 (Primary) 68T10; 68T37 (Secondary) ACM Class: I.2.6; I.5.1; I.5.4; J.2

On Orbits and the Finiteness of Bounded Automaton Groups

Authors: Ievgen Bondarenko, Jan Philipp Wächter

Abstract: We devise an algorithm which, given a bounded automaton A, decides whether the group generated by A is finite. The solution comes from a description of the infinite sequences having an infinite A-orbit using a deterministic finite-state acceptor. This acceptor can also be used to decide whether the bounded automaton acts level-transitively. We devise an algorithm which, given a bounded automaton A, decides whether the group generated by A is finite. The solution comes from a description of the infinite sequences having an infinite A-orbit using a deterministic finite-state acceptor. This acceptor can also be used to decide whether the bounded automaton acts level-transitively. △ Less

Submitted 8 September, 2021; v1 submitted 14 December, 2019; originally announced December 2019.

Comments: Content and theorem numbers match the published version

MSC Class: 20F10; 20M35 ACM Class: F.4.m

Journal ref: International Journal of Algebra and Computation, Vol. 31, No. 06, pp. 1177-1190 (2021)

The lamplighter group $\mathbb{Z}_3\wr\mathbb{Z}$ generated by a bireversible automaton

Authors: I. Bondarenko, D. D'Angeli, E. Rodaro

Abstract: We construct a bireversible self-dual automaton with $3$ states over an alphabet with $3$ letters which generates the lamplighter group $\mathbb{Z}_3\wr\mathbb{Z}$. We construct a bireversible self-dual automaton with $3$ states over an alphabet with $3$ letters which generates the lamplighter group $\mathbb{Z}_3\wr\mathbb{Z}$. △ Less

Submitted 27 February, 2015; originally announced February 2015.

The word problem in Hanoi Towers groups

Authors: Ievgen Bondarenko

Abstract: We prove that elements of the Hanoi Towers groups $\mathcal{H}_m$ have depth bounded from above by a poly-logarithmic function $O(\log^{m-2} n)$, where $n$ is the length of an element. Therefore the word problem in groups $\mathcal{H}_m$ is solvable in subexponential time $\exp(O(\log^{m-2} n))$. We prove that elements of the Hanoi Towers groups $\mathcal{H}_m$ have depth bounded from above by a poly-logarithmic function $O(\log^{m-2} n)$, where $n$ is the length of an element. Therefore the word problem in groups $\mathcal{H}_m$ is solvable in subexponential time $\exp(O(\log^{m-2} n))$. △ Less

Submitted 30 August, 2014; originally announced September 2014.

MSC Class: 20F10; 68R05