Zobrazeno 1 - 10
of 36
pro vyhledávání: '"Hązła, Jan"'
Autor:
Alidou, Abdou Majeed, Baligács, Júlia, Hahn-Klimroth, Max, Hązła, Jan, Hintze, Lukas, Scheftelowitsch, Olga
Polarization and unexpected correlations between opinions on diverse topics (including in politics, culture and consumer choices) are an object of sustained attention. However, numerous theoretical models do not seem to convincingly explain these phe
Externí odkaz:
http://arxiv.org/abs/2402.08446
Autor:
Kougang-Yombi, Donald, Hązła, Jan
This paper introduces a quantitative generalization of the ``more capable'' comparison of broadcast channels, which is termed ``more capable with advantage''. Some basic properties are demonstrated (including tensorization on product channels), and a
Externí odkaz:
http://arxiv.org/abs/2401.14214
Autor:
Djagba, Prudence, Hązła, Jan
Combinatorial aspects of R-subgroups of finite dimensional Beidleman near-vector spaces over nearfields are studied. A characterization of R-subgroups is used to obtain the smallest possible size of a generating set of a subgroup, which is much small
Externí odkaz:
http://arxiv.org/abs/2306.16421
Autor:
Hązła, Jan
A noisy entropy inequality for boolean functions by Samorodnitsky is applied to binary codes. It is shown that a binary code that achieves capacity on the binary erasure channel admits optimal list size for list decoding on some binary symmetric chan
Externí odkaz:
http://arxiv.org/abs/2212.01443
Publikováno v:
Proceedings of the International Conference on Machine Learning, 2022
This paper introduces the notion of ``Initial Alignment'' (INAL) between a neural network at initialization and a target function. It is proved that if a network and a Boolean target function do not have a noticeable INAL, then noisy gradient descent
Externí odkaz:
http://arxiv.org/abs/2202.12846
How does the geometric representation of a dataset change after the application of each randomly initialized layer of a neural network? The celebrated Johnson--Lindenstrauss lemma answers this question for linear fully-connected neural networks (FNNs
Externí odkaz:
http://arxiv.org/abs/2111.02155
We study the implicit bias of ReLU neural networks trained by a variant of SGD where at each step, the label is changed with probability $p$ to a random label (label smoothing being a close variant of this procedure). Our experiments demonstrate that
Externí odkaz:
http://arxiv.org/abs/2111.02154
Autor:
Cornacchia, Elisabetta, Hązła, Jan
We settle a version of the conjecture about intransitive dice posed by Conrey, Gabbard, Grant, Liu and Morrison in 2016 and Polymath in 2017. We consider generalized dice with $n$ faces and we say that a die $A$ beats $B$ if a random face of $A$ is m
Externí odkaz:
http://arxiv.org/abs/2011.10067
This paper considers '$\delta$-almost Reed-Muller codes', i.e., linear codes spanned by evaluations of all but a $\delta$ fraction of monomials of degree at most $d$. It is shown that for any $\delta > 0$ and any $\varepsilon>0$, there exists a famil
Externí odkaz:
http://arxiv.org/abs/2004.09590
We introduce a simple, geometric model of opinion polarization. It is a model of political persuasion, as well as marketing and advertising, utilizing social values. It focuses on the interplay between different topics and persuasion efforts. We demo
Externí odkaz:
http://arxiv.org/abs/1910.05274