Zobrazeno 1 - 10
of 84
pro vyhledávání: '"DARÓCZY, Bálint"'
Many state-of-the-art models trained on long-range sequences, for example S4, S5 or LRU, are made of sequential blocks combining State-Space Models (SSMs) with neural networks. In this paper we provide a PAC bound that holds for these kind of archite
Externí odkaz:
http://arxiv.org/abs/2405.20278
One of the main theoretical challenges in learning dynamical systems from data is providing upper bounds on the generalization error, that is, the difference between the expected prediction error and the empirical prediction error measured on some fi
Externí odkaz:
http://arxiv.org/abs/2405.10054
Recent advances in deep learning have given us some very promising results on the generalization ability of deep neural networks, however literature still lacks a comprehensive theory explaining why heavily over-parametrized models are able to genera
Externí odkaz:
http://arxiv.org/abs/2310.17378
Autor:
Pál, Dániel Levente, Daróczy, Bálint
Publikováno v:
MatLit, Vol 4, Iss 2, Pp 77-97 (2016)
Using an abstract graph model we describe a hypothetical transformation of a literary entity through the active connections in the graph. First, we weakly define a set of transformations (“morphology”) over a particular entity as a series of the
Externí odkaz:
https://doaj.org/article/4880ef1588db431f8d6f2ab9fb388977
We consider the problem of learning Neural Ordinary Differential Equations (neural ODEs) within the context of Linear Parameter-Varying (LPV) systems in continuous-time. LPV systems contain bilinear systems which are known to be universal approximato
Externí odkaz:
http://arxiv.org/abs/2307.03630
Autor:
Rácz, Dániel, Daróczy, Bálint
Feed-forward networks can be interpreted as mappings with linear decision surfaces at the level of the last layer. We investigate how the tangent space of the network can be exploited to refine the decision in case of ReLU (Rectified Linear Unit) act
Externí odkaz:
http://arxiv.org/abs/2110.13581
Building on the quantum ensemble based classifier algorithm of Schuld and Petruccione [arXiv:1704.02146v1], we devise equivalent classical algorithms which show that this quantum ensemble method does not have advantage over classical algorithms. Esse
Externí odkaz:
http://arxiv.org/abs/2102.00949
Autor:
Daróczy, Bálint
Recent articles indicate that deep neural networks are efficient models for various learning problems. However they are often highly sensitive to various changes that cannot be detected by an independent observer. As our understanding of deep neural
Externí odkaz:
http://arxiv.org/abs/2006.06780
Hierarchical neural networks are exponentially more efficient than their corresponding "shallow" counterpart with the same expressive power, but involve huge number of parameters and require tedious amounts of training. By approximating the tangent s
Externí odkaz:
http://arxiv.org/abs/1912.09306
Hierarchical neural networks are exponentially more efficient than their corresponding "shallow" counterpart with the same expressive power, but involve huge number of parameters and require tedious amounts of training. Our main idea is to mathematic
Externí odkaz:
http://arxiv.org/abs/1807.06630