Zobrazeno 1 - 10
of 981
pro vyhledávání: '"A P, Anokhin"'
Autor:
Kuratov, Yuri, Bulatov, Aydar, Anokhin, Petr, Rodkin, Ivan, Sorokin, Dmitry, Sorokin, Artyom, Burtsev, Mikhail
In recent years, the input context sizes of large language models (LLMs) have increased dramatically. However, existing evaluation methods have not kept pace, failing to comprehensively assess the efficiency of models in handling long contexts. To br
Externí odkaz:
http://arxiv.org/abs/2406.10149
Autor:
Kuratov, Yuri, Bulatov, Aydar, Anokhin, Petr, Sorokin, Dmitry, Sorokin, Artyom, Burtsev, Mikhail
This paper addresses the challenge of processing long documents using generative transformer models. To evaluate different approaches, we introduce BABILong, a new benchmark designed to assess model capabilities in extracting and processing distribut
Externí odkaz:
http://arxiv.org/abs/2402.10790
We study a toy-model of continuous infinite expansion of space-time with the flat start. We use as the gravitational background a conformaly flat metric with an exponentially growing factor in conformal time. We aim to clarify some properties of quan
Externí odkaz:
http://arxiv.org/abs/2401.12855
Elements of neural networks, both biological and artificial, can be described by their selectivity for specific cognitive features. Understanding these features is important for understanding the inner workings of neural networks. For a living system
Externí odkaz:
http://arxiv.org/abs/2401.10748
Publikováno v:
Âderna Fìzika ta Energetika, Vol 25, Iss 3, Pp 241-250 (2024)
The absorption cross-section of Mössbauer radiation in magnetic liquids is calculated, taking into consideration both translational and rotational Brownian motion of magnetic nanoparticles as well as stochastic reversals of their magnetization in th
Externí odkaz:
https://doaj.org/article/e5e668cb5a894a8ba6a392289337ef64
We propose the Thinker algorithm, a novel approach that enables reinforcement learning agents to autonomously interact with and utilize a learned world model. The Thinker algorithm wraps the environment with a world model and introduces new actions d
Externí odkaz:
http://arxiv.org/abs/2307.14993
Autor:
Sorokin, Konstantin, Ayzenberg, Anton, Anokhin, Konstantin, Sotskov, Vladimir, Beketov, Maxim, Zaitsew, Andrey, Drynkin, Robert
In present paper we discuss several approaches to reconstructing the topology of the physical space from neural activity data of CA1 fields in mice hippocampus, in particular, having Cognitome theory of brain function in mind. In our experiments, ani
Externí odkaz:
http://arxiv.org/abs/2211.08718
Publikováno v:
Известия высших учебных заведений. Поволжский регион: Медицинские науки, Iss 2 (2024)
Background. Atypical hemolytic uremic syndrome is an ultra-rare (orphan) disease from the group of thrombotic microangiopathies of progressive course, which is caused by uncontrolled activation of the alternative complement pathway of hereditary or
Externí odkaz:
https://doaj.org/article/2d6a6fe629d14dec80f5acf1658bf501
Publikováno v:
Frontiers in Pharmacology, Vol 15 (2024)
Introduction: Reactivation of already consolidated memory can initiate its destabilization, making the memory trace labile. Normally, this destabilization is followed by reconsolidation of the memory trace, enriched by newly acquired experience. Disr
Externí odkaz:
https://doaj.org/article/2ae8de715bdc489d85d0ec810383012a
Autor:
Velikanov, Maksim, Kail, Roman, Anokhin, Ivan, Vashurin, Roman, Panov, Maxim, Zaytsev, Alexey, Yarotsky, Dmitry
A memory efficient approach to ensembling neural networks is to share most weights among the ensembled models by means of a single reference network. We refer to this strategy as Embedded Ensembling (EE); its particular examples are BatchEnsembles an
Externí odkaz:
http://arxiv.org/abs/2202.12297