检索结果-内蒙古大学图书馆

Using Machine Learning Techniques for Assessing Media Text Readability 24

学校读者我要写书评

暂无评论

Using Machine Learning Techniques for Assessing Media Text R...

24th IEEE International Conference of Young Professionals in Electron Devices and Materials, EDM 2023

作者： Laputenko, Andrey Evtushenko, Tatiana Klochkova, Elena Yevtushenko, Nina National Research Tomsk State University Department of Information Technologies Tomsk Russia Graduate School of Applied Linguistics and Translation Studies Peter the Great St. Petersburg Polytechnic University Saint Petersburg Russia Ivannikov Institute for System Programming of the Russian Academy of Sciences Department of Software Engineering Moscow Russia

ISBN: (纸本)9798350336870

The paper addresses the problem of assessing media text readability depending on the correlation of objective and subjective text complexity. Objective text complexity refers to a set of measurable characteristics such as baseline, morphological, syntactic and lexical characteristics of a media text (a piece of news published on university websites), which is calculated automatically for each text. In total, 34 parameters are considered. Subjective text complexity score is assigned to each text by human experts who assess the clarity, structure, cohesion and coherence of the text. The correlation between subjective and objective text complexity is studied using machine learning models. The readability level is measured using machine learning methods and regression-correlation analysis. Namely, an artificial neural network and regression models are used. It is demonstrated that the use of polynomial features and lasso regularization allows to get a compact regression model of a high quality. The use of the chosen machine learning techniques made it possible to estimate the impact of distinct linguistic features on the readability level. © 2023 IEEE.

关键词： Machine learning

Finetuning BERT on Partially Annotated NER Corpora

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Scherbakov, Viktor Mayorov, Vladimir Ivannikov Institute for System Programming The Russian Academy of Sciences Russia Lomonosov Moscow State University Russia

Most Named Entity Recognition (NER) models operate under the assumption that training datasets are fully labelled. While it is valid for established datasets like CoNLL 2003 and OntoNotes, sometimes it is not feasible to obtain the complete dataset annotation. These situations may occur, for instance, after selective annotation of entities for cost reduction. This work presents an approach to finetuning BERT on such partially labelled datasets using self-supervision and label preprocessing. Our approach outperforms the previous LSTM-based label preprocessing baseline, significantly improving the performance on poorly labelled datasets. We demonstrate that following our approach while finetuning RoBERTa on CoNLL 2003 dataset with only 10% of total entities labelled is enough to reach the performance of the baseline trained on the same dataset with 50% of the entities labelled. © 2022, CC BY-SA.

关键词： Cost reduction

Deep Neural Networks Generalization and Fine-Tuning for 12-lead ECG Classification

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Avetisyan, Aram Tigranyan, Shahane Asatryan, Ariana Mashkova, Olga Skorik, Sergey Ananev, Vladislav Markin, Yury The Ivannikov Institute for System Programming The Russian Academy of Sciences Moscow109004 Russia The Russian-Armenian University Yerevan375051 Armenia The Novgorod State University Veliky Novgorod173003 Russia

Numerous studies are aimed at diagnosing heart diseases based on 12-lead electrocardiographic (ECG) records using deep learning methods. These studies usually use specific datasets that differ in size and parameters, such as patient metadata, number of doctors annotating ECGs, types of devices for ECG recording, data preprocessing techniques, etc. It is well-known that high-quality deep neural networks trained on one ECG dataset do not necessarily perform well on another dataset or clinical settings. In this paper, we propose a methodology to improve the quality of heart disease prediction regardless of the dataset by training neural networks on a variety of datasets with further fine-tuning for the specific dataset. To show its applicability, we train different neural networks on a large private dataset TIS containing various ECG records from multiple hospitals and on a relatively small public dataset PTB-XL. We demonstrate that training the networks on a large dataset and fine-tuning it on a small dataset from another source outperforms the networks trained only on one small dataset. We also show how the ability of a deep neural networks to generalize allows to improve classification quality of more diseases. Copyright © 2023, The Authors. All rights reserved.

关键词： Electrocardiography

Classifiers of the Functional State of Living systems Based on Voigt Model Optimization 2nd

学校读者我要写书评

暂无评论

Classifiers of the Functional State of Living Systems Based ...

2nd International Conference on Applied Intelligence, ICAI 2024

作者： Al-kasasbeh, Riad Taha Filist, Sergey Shaqadan, Ashraf Shatalova, Olga Starkov, Evgeny Korenevskiy, Nikolay A. Al-Habahbeh, Osama M. Tarawneh, Emad Saleh Maksim, Ilyash Miroshnikova, Karina Miroshnikov, Andrey Aikeyeva, Altyn A. The University of Jordan Amman Jordan South-West State University Kursk Russia Zarqa University Zarqa Jordan Institute of System Programming of the Russian Academy of Sciences Moscow Russia Eurasian National University Astana Kazakhstan

ISBN: (纸本)9789819619061

A novel bioimpedance spectroscopy method is introduced;the proposed method can help in the development of objective and accessible criteria for disease assessment, in addition to evaluating the effectiveness of treatment methods. The proposed method can be utilized to develop criteria for conservative therapy options and surgical interventions in severe cases. The method employs a recurrent Voigt model to represent biomaterial segment impedance. For each biomaterial segment model, a Cole plot is drawn in a given frequency range. The model parameters allow the formation of descriptors for multimodal classifiers of the functional state of living systems and parameters of the model links. To confirm the validity of the classifier, a group of patients diagnosed with pneumonia have been tested. To obtain bioimpedance analysis data, an electrode belt was placed on the chest of each patient, and impedance diagrams corresponding to a certain combination of electrodes were determined. The quality indicators of various classifier models reached 78% and were above 62%. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Diseases

On greedy approximation in complex Banach spaces

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Gasnikov, A. Temlyakov, V. Ivannikov institute for System Programming Russian Academy of Sciences Moscow Russia Steklov Mathematical Institute Russian Academy of Sciences Moscow Russia Innopolis University Tatarstan Russia University of South Carolina United States Lomonosov Moscow State University Russia Moscow Center of Fundamental and Applied Mathematics Russia

The general theory of greedy approximation with respect to arbitrary dictionaries is well developed in the case of real Banach spaces. Recently, some of results proved for the Weak Chebyshev Greedy Algorithm (WCGA) in the case of real Banach spaces were extended to the case of complex Banach spaces. In this paper we extend some of known in the real case results for other than WCGA greedy algorithms to the case of complex Banach spaces. Copyright © 2024, The Authors. All rights reserved.

关键词： Banach spaces

Decentralised convex optimisation with probability-proportional-to-size quantization

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Pasechnyuk, D.A. Dvurechensky, P. Uribe, C.A. Gasnikov, A.V. Mohamed bin Zayed University of Artificial Intelligence United Arab Emirates Ivannikov Institute for System Programming the Russian Academy of Sciences Russia Weierstrass Institute for Applied Analysis and Stochastics Germany Rice University United States Moscow Institute of Physics and Technology Russia

Communication is one of the bottlenecks of distributed optimisation and learning. To overcome this bottleneck, we propose a novel quantization method that transforms a vector into a sample of components’ indices drawn from a categorical distribution with probabilities proportional to values at those components. Then, we propose a primal and a primal-dual accelerated stochastic gradient methods that use our proposed quantization, and derive their convergence rates in terms of probabilities of large deviations. We focus on affine-constrained convex optimisation and its application to decentralised distributed optimisation problems. To illustrate the work of our algorithm, we apply it to the decentralised computation of semi-discrete entropy regularized Wasserstein *** Codes 65K10 © 2025, CC BY.

关键词： Optimization algorithms

A Dataset for Information Extraction from News Web Pages

学校读者我要写书评

暂无评论

A Dataset for Information Extraction from News Web Pages

Ivannikov ISPRAS Open Conference (ISPRAS)

作者： Maksim Varlamov Denis Galanin Pavel Bedrin Sergey Duda Vladimir Lazarev Alexander Yatskov Ivannikov Institute for System Programming of the Russian Academy of Sciences Moscow Russia Lomonosov Moscow State University Moscow Russia

In this paper we propose a new dataset for information extraction from news web pages. Accurate collection of news articles is necessary to build systems that aggregate and analyze data from a large number of news sources. However, existing news page datasets are designed for the problem of article content extraction and do not consider other metadata such as dates or authors, while modern generic approaches for structured web data extraction are evaluated on other domains, mostly e-commerce products websites. Our dataset contains 724 web pages from 114 russian news websites. On each page we manually annotated title, text, publication date, tags and other article attributes. We describe data collection and annotation process and demonstrate evaluation results for open source tools and neural network based approaches. The dataset is publicly available on GitHub.

关键词： Annotations Neural networks Web pages Data aggregation Information retrieval Data mining Electronic commerce

Applying language models to algebraic topology: generating simplicial cycles using multi-labeling in Wu's formula 24

学校读者我要写书评

暂无评论

Applying language models to algebraic topology: generating s...

Proceedings of the 41st International Conference on Machine Learning

作者： Kirill Brilliantov Fedor Pavutnitskiy Dmitry Pasechnyuk German Magai ETH Zurich Beijing Institute of Mathematical Sciences and Applications Beijing Mohamed bin Zayed University of Artificial Intelligence Abu Dhabi and Ivannikov Institute for System Programming of the Russian Academy of Sciences Moscow HSE University Moscow and Noeon Research Tokyo

Computing homotopy groups of spheres has long been a fundamental objective in algebraic topology. Various theoretical and algorithmic approaches have been developed to tackle this problem. In this paper we take a step towards the goal of comprehending the group-theoretic structure of the generators of these homotopy groups by leveraging the power of machine learning. Specifically, in the simplicial group setting of Wu's formula, we reformulate the problem of generating simplicial cycles as a problem of sampling from the intersection of algorithmic datasets related to Dyck languages. We present and evaluate language modelling approaches that employ multilabel information for input sequences, along with the necessary group-theoretic toolkit and nonneural baselines.

关键词：

Analytical and numerical methods for Zhukovsky airfoils aerodynamics coefficients

学校读者我要写书评

暂无评论

Analytical and numerical methods for Zhukovsky airfoils aero...

Ivannikov ISPRAS Open Conference (ISPRAS)

作者： A.G. Petrov A.D. Sukhov I.N. Sibgatullin A.D. Britov Ishlinsky Institute for Problems in Mechanics RAS Moscow 119526 Russia Institute for System Programming of the Russian Academy of Sciences Moscow 109004 Russia

The two-dimensional problem of a viscous laminar flow around Zhukovsky airfoils at an angle of attack is considered. Based on the approach of local similarity, which was proposed by Kochin and Loytsyansky for the equations of laminar boundary layer, we have found the shear stresses at the aifoil, and the coordinates of the separation points. Assuming the values of the velocities at the separation points to be equal, we find the value of the circulation. A complete solution to the problem of the velocity and pressure field outside the boundary layer is also constructed. The theoretical results are compared with the available experimental data and numerical simulations of the Navier-Stokes equations.

关键词： Drag Force Numerical simulation Aerodynamics Mathematical models Automotive components Stress