<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">ntv</journal-id><journal-title-group><journal-title xml:lang="ru">Научно-технический вестник информационных технологий, механики и оптики</journal-title><trans-title-group xml:lang="en"><trans-title>Scientific and Technical Journal of Information Technologies, Mechanics and Optics</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2226-1494</issn><issn pub-type="epub">2500-0373</issn><publisher><publisher-name>Университет ИТМО</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.17586/2226-1494-2023-23-6-1187-1197</article-id><article-id custom-type="elpub" pub-id-type="custom">ntv-33</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>ИСКУССТВЕННЫЙ ИНТЕЛЛЕКТ И КОГНИТИВНЫЕ ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>ARTIFICIAL INTELLIGENCE AND COGNITIVE INFORMATION TECHNOLOGIES</subject></subj-group></article-categories><title-group><article-title>Использование топологического анализа данных для построения байесовских нейронных сетей</article-title><trans-title-group xml:lang="en"><trans-title>Using topological data analysis for building Bayesan neural networks</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0002-5483-716X</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Ватьян</surname><given-names>А. С.</given-names></name><name name-style="western" xml:lang="en"><surname>Vatian</surname><given-names>A. S.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Ватьян Александра Сергеевна — кандидат технических наук, доцент</p><p>Санкт-Петербург, 197101</p><p>sc 57191870868</p></bio><bio xml:lang="en"><p>Alexandra S. Vatian — PhD, Associate Professor</p><p>Saint Petersburg, 197101</p><p>sc 57191870868</p></bio><email xlink:type="simple">alexvatyan@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Гусарова</surname><given-names>Н. Ф.</given-names></name><name name-style="western" xml:lang="en"><surname>Gusarova</surname><given-names>N. F.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Гусарова Наталия Федоровна — кандидат технических наук, доцент</p><p>Санкт-Петербург, 197101</p><p>sc 57162764200</p></bio><bio xml:lang="en"><p>Natalia F. Gusarova — PhD, Associate Professor</p><p>Saint Petersburg, 197101</p><p>sc 57162764200</p></bio><email xlink:type="simple">natfed@list.ru</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0006-1485-1166</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Добренко</surname><given-names>Д. А.</given-names></name><name name-style="western" xml:lang="en"><surname>Dobrenko</surname><given-names>D. A.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Добренко Дмитрий Александрович — аспирант</p><p>Санкт-Петербург, 197101</p></bio><bio xml:lang="en"><p>Dmitriy A. Dobrenko — PhD Student</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">enotpalaskun@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0002-0490-0148</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Панкова</surname><given-names>К. С.</given-names></name><name name-style="western" xml:lang="en"><surname>Pankova</surname><given-names>K. S.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Панкова Кристина Сергеевна — студент</p><p>Санкт-Петербург, 197101</p></bio><bio xml:lang="en"><p>Kristina S. Pankova — Student</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">pankovakrisss@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0003-1886-2867</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Томилов</surname><given-names>И. В.</given-names></name><name name-style="western" xml:lang="en"><surname>Tomilov</surname><given-names>I. V.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Томилов Иван Вячеславович — старший лаборант</p><p>Санкт-Петербург, 197101</p><p>sc 57772599000</p></bio><bio xml:lang="en"><p>Ivan V. Tomilov — Senior Laboratory Assistant</p><p>sc 57772599000</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">ivan-tomilov3@yandex.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>Университет ИТМО</institution><country>Россия</country></aff><aff xml:lang="en"><institution>ITMO University</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2023</year></pub-date><pub-date pub-type="epub"><day>11</day><month>12</month><year>2024</year></pub-date><volume>23</volume><issue>6</issue><fpage>1187</fpage><lpage>1197</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Ватьян А.С., Гусарова Н.Ф., Добренко Д.А., Панкова К.С., Томилов И.В., 2024</copyright-statement><copyright-year>2024</copyright-year><copyright-holder xml:lang="ru">Ватьян А.С., Гусарова Н.Ф., Добренко Д.А., Панкова К.С., Томилов И.В.</copyright-holder><copyright-holder xml:lang="en">Vatian A.S., Gusarova N.F., Dobrenko D.A., Pankova K.S., Tomilov I.V.</copyright-holder><license xml:lang="ru" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>Данная работа распространяется под лицензией Creative Commons Attribution 4.0.</license-p></license><license xml:lang="en" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://ntv.elpub.ru/jour/article/view/33">https://ntv.elpub.ru/jour/article/view/33</self-uri><abstract><p>Введение. Впервые предложен упрощенный подход к построению байесовских нейронных сетей, сочетающий вычислительную эффективность с возможностью анализа процесса обучения. Метод. Предлагаемый подход основан на байесианизации детерминированной нейронной сети посредством рандомизации параметров только на уровне интерфейса. Формирование байесовской нейронной сети на основе заданной сети осуществляется путем замены ее параметров на вероятностные распределения, которые имеют в качестве среднего значения параметры исходной модели. Оценки метрик эффективности нейронной сети, полученной в рамках рассматриваемого подхода, и байесовской нейронной сети, построенной посредством вариационного вывода, выполнены с использованием методов топологического анализа данных. Основные результаты. Процедура байесианизации реализована с помощью градуированного варьирования интенсивности рандомизации. В качестве альтернативы использованы две нейронные сети с идентичной структурой — детерминированная и классическая байесовская. На вход нейронной сети подавались исходные данные двух датасетов из медицинского домена в вариантах без зашумления и с добавленным гауссовским шумом. Рассчитаны нулевые и первые персистентные гомологии для эмбеддингов формируемых нейронных сетей на каждом из слоев. Для оценки качества классификации использована метрика точности (accuracy). Показано, что баркоды для эмбеддингов на каждом слое байесианизированной нейронной сети во всех четырех сценариях находятся между соответствующими баркодами детерминированной и байесовской нейронной сетей как для нулевых, так и для первых персистентных гомологий. При этом детерминированная нейронная сеть является нижней границей, а байесовская — верхней. Показано, что структура ассоциаций данных внутри байесианизированной нейронной сети наследуется от детерминированной модели, однако приобретает свойства байесовской. Экспериментально установлено наличие взаимосвязи между нормированной персистентной энтропией, вычисляемой на эмбеддингах нейронной сети, и точностью нейронной сети. Для предсказания точности наиболее показательной оказалась топология эмбеддингов на среднем слое модели нейронной сети. Обсуждение. Предлагаемый подход может быть использован для упрощения построения байесовской нейронной сети из уже обученной детерминированной нейронной сети. Это открывает возможности повышения точности существующей нейронной сети без ансамблирования с дополнительными классификаторами. Появляется возможность проактивной оценки эффективности формируемой нейронной сети на упрощенных данных без запуска на реальном датасете, что сокращает ресурсоемкость ее разработки.</p></abstract><trans-abstract xml:lang="en"><p>For the first time, a simplified approach to constructing Bayesian neural networks is proposed, combining computational efficiency with the ability to analyze the learning process. The proposed approach is based on Bayesianization of a deterministic neural network by randomizing parameters only at the interface level, i.e., the formation of a Bayesian neural network based on a given network by replacing its parameters with probability distributions that have the parameters of the original model as the average value. Evaluations of the efficiency metrics of the neural network were obtained within the framework of the approach under consideration, and the Bayesian neural network constructed through variation inference were performed using topological data analysis methods. The Bayesianization procedure is implemented through graded variation of the randomization intensity. As an alternative, two neural networks with identical structure were used — deterministic and classical Bayesian networks. The input of the neural network was supplied with the original data of two datasets in versions without noise and with added Gaussian noise. The zero and first persistent homologies for the embeddings of the formed neural networks on each layer were calculated. To assess the quality of classification, the accuracy metric was used. It is shown that the barcodes for embeddings on each layer of the Bayesianized neural network in all four scenarios are between the corresponding barcodes of the deterministic and Bayesian neural networks for both zero and first persistent homologies. In this case, the deterministic neural network is the lower bound, and the Bayesian neural network is the upper bound. It is shown that the structure of data associations within a Bayesianized neural network is inherited from a deterministic model, but acquires the properties of a Bayesian one. It has been experimentally established that there is a relationship between the normalized persistent entropy calculated on neural network embeddings and the accuracy of the neural network. For predicting accuracy, the topology of embeddings on the middle layer of the neural network model turned out to be the most revealing. The proposed approach can be used to simplify the construction of a Bayesian neural network from an already trained deterministic neural network, which opens up the possibility of increasing the accuracy of an existing neural network without ensemble with additional classifiers. It becomes possible to proactively evaluate the effectiveness of the generated neural network on simplified data without running it on a real dataset, which reduces the resource intensity of its development.</p></trans-abstract><kwd-group xml:lang="ru"><kwd>байесовские нейронные сети</kwd><kwd>персистентная гомология</kwd><kwd>нормированная персистентная энтропия</kwd><kwd>эмбеддинг</kwd><kwd>баркод</kwd></kwd-group><kwd-group xml:lang="en"><kwd>Bayesian neural networks</kwd><kwd>persistent homology</kwd><kwd>normalized persistent entropy</kwd><kwd>embedding</kwd><kwd>barcode</kwd></kwd-group><funding-group><funding-statement xml:lang="ru">Работа поддержана грантом Российского научного фонда 23-11-00346.</funding-statement><funding-statement xml:lang="en">The work is supported by Grant RSF 23-11-00346.</funding-statement></funding-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Chazal F., Michel B. An introduction to topological data analysis: fundamental and practical aspects for data scientists // Frontiers in Artificial Intelligence. 2021. V. 4. https://doi.org/10.3389/frai.2021.667963</mixed-citation><mixed-citation xml:lang="en">Chazal F., Michel B. An introduction to topological data analysis: fundamental and practical aspects for data scientists. Frontiers in Artificial Intelligence, 2021, vol. 4. https://doi.org/10.3389/frai.2021.667963</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Edelsbrunner H., Harer J. Computational topology: an introduction. American Mathe-matical Soc., 2010 [Электронный ресурс]. URL: https://www.maths.ed.ac.uk/~v1ranick/papers/edelcomp.pdf (дата обращения: 10.11.2023).</mixed-citation><mixed-citation xml:lang="en">Edelsbrunner H., Harer J. Computational topology: an introduction. American Mathe-matical Soc., 2010. Available at: https://www.maths.ed.ac.uk/~v1ranick/papers/edelcomp.pdf (accessed: 10.11.2023).</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Ritter H., Kukla M., Zhang C., Li Y. Sparse uncertainty representation in deep learning with inducing weights // Advances in Neural Information Processing Systems. 2021. V. 8. P. 6515–6528.</mixed-citation><mixed-citation xml:lang="en">Ritter H., Kukla M., Zhang C., Li Y. Sparse uncertainty representation in deep learning with inducing weights. Advances in Neural Information Processing Systems, 2021, vol. 8, pp. 6515–6528.</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">Prabhudesai S., Hauth J., Guo D., Rao A., Banovic N., Huan X. Lowering the computational barrier: Partially Bayesian neural networks for transparency in medical imaging AI // Frontiers in Computer Science. 2023. V. 5. https://doi.org/10.3389/fcomp.2023.1071174</mixed-citation><mixed-citation xml:lang="en">Prabhudesai S., Hauth J., Guo D., Rao A., Banovic N., Huan X. Lowering the computational barrier: Partially Bayesian neural networks for transparency in medical imaging AI. Frontiers in Computer Science, 2023, vol. 5. https://doi.org/10.3389/fcomp.2023.1071174</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Zomorodian A., Carlsson G. Computing persistent homology // Discrete &amp; Computational Geometry. 2005. V. 33. N 2. P. 249–274. https://doi.org/10.1007/s00454-004-1146-y</mixed-citation><mixed-citation xml:lang="en">Zomorodian A., Carlsson G. Computing persistent homology. Discrete &amp; Computational Geometry, 2005, vol. 33, no. 2, pp. 249– 274. https://doi.org/10.1007/s00454-004-1146-y</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Wasserman L. Topological data analysis // Annual Review of Statistics and Its Application. 2018. V. 5. P. 501–532. https://doi.org/10.1146/annurev-statistics-031017-100045</mixed-citation><mixed-citation xml:lang="en">Wasserman L. Topological data analysis. Annual Review of Statistics and Its Application, 2018, vol. 5, pp. 501–532. https://doi.org/10.1146/annurev-statistics-031017-100045</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Carlsson G., Gabrielsson R.B. Topological approaches to deep learning // Topological Data Analysis. Springer, 2020. P. 119–146. https://doi.org/10.1007/978-3-030-43408-3_5</mixed-citation><mixed-citation xml:lang="en">Carlsson G., Gabrielsson R.B. Topological approaches to deep learning. Topological Data Analysis. Springer, 2020, pp. 119–146. https://doi.org/10.1007/978-3-030-43408-3_5</mixed-citation></citation-alternatives></ref><ref id="cit8"><label>8</label><citation-alternatives><mixed-citation xml:lang="ru">Hensel F., Moor M., Rieck B. A survey of topological machine learning methods // Frontiers in Artificial Intelligence. 2021. V. 4. https://doi.org/10.3389/frai.2021.681108</mixed-citation><mixed-citation xml:lang="en">Hensel F., Moor M., Rieck B. A survey of topological machine learning methods. Frontiers in Artificial Intelligence, 2021, vol. 4. https://doi.org/10.3389/frai.2021.681108</mixed-citation></citation-alternatives></ref><ref id="cit9"><label>9</label><citation-alternatives><mixed-citation xml:lang="ru">Moroni D., Pascali M.A. Learning topology: bridging computational topology and machine learning // Pattern Recognition and Image Analysis. 2021. V. 31. N 3. P. 443–453. https://doi.org/10.1134/S1054661821030184</mixed-citation><mixed-citation xml:lang="en">Moroni D., Pascali M.A. Learning topology: bridging computational topology and machine learning. Pattern Recognition and Image Analysis, 2021, vol. 31, no. 3, pp. 443–453. https://doi.org/10.1134/S1054661821030184</mixed-citation></citation-alternatives></ref><ref id="cit10"><label>10</label><citation-alternatives><mixed-citation xml:lang="ru">Zia A., Khamis A., Nichols J., Hayder Z., Rolland V., Peterssonet L. Topological deep learning: A review of an emerging paradigm // arXiv. arXiv:2302.03836v1. 2023. https://doi.org/10.48550/arXiv.2302.03836</mixed-citation><mixed-citation xml:lang="en">Zia A., Khamis A., Nichols J., Hayder Z., Rolland V., Peterssonet L. Topological deep learning: A review of an emerging paradigm. arXiv, arXiv:2302.03836v1, 2023. https://doi.org/10.48550/arXiv.2302.03836</mixed-citation></citation-alternatives></ref><ref id="cit11"><label>11</label><citation-alternatives><mixed-citation xml:lang="ru">Goibert M., Ricatte T., Dohmatob E. An adversarial robustness perspective on the topology of neural networks // arXiv. 2022. arXiv:2211.02675. https://doi.org/10.48550/arXiv.2211.02675</mixed-citation><mixed-citation xml:lang="en">Goibert M., Ricatte T., Dohmatob E. An adversarial robustness perspective on the topology of neural networks. arXiv, 2022, arXiv:2211.02675. https://doi.org/10.48550/arXiv.2211.02675</mixed-citation></citation-alternatives></ref><ref id="cit12"><label>12</label><citation-alternatives><mixed-citation xml:lang="ru">Chen C., Ni X., Bai Q., Wang Y. A topological regularizer for classifiers via persistent homology // Proc. of the AISTATS 2019 — 22nd International Conference on Artificial Intelligence and Statistics. 2020.</mixed-citation><mixed-citation xml:lang="en">Chen C., Ni X., Bai Q., Wang Y. A topological regularizer for classifiers via persistent homology. Proc. of the AISTATS 2019 — 22nd International Conference on Artificial Intelligence and Statistics, 2020.</mixed-citation></citation-alternatives></ref><ref id="cit13"><label>13</label><citation-alternatives><mixed-citation xml:lang="ru">Ramamurthy K.N., Varshney K.R., Mody K. Topological data analysis of decision boundaries with application to model selection // Proc. of the 36th International Conference on Machine Learning (ICML). 2019. P. 9316–9325.</mixed-citation><mixed-citation xml:lang="en">Ramamurthy K.N., Varshney K.R., Mody K. Topological data analysis of decision boundaries with application to model selection. Proc. of the 36th International Conference on Machine Learning (ICML), 2019, pp. 9316–9325.</mixed-citation></citation-alternatives></ref><ref id="cit14"><label>14</label><citation-alternatives><mixed-citation xml:lang="ru">Gabrielsson R.B., Carlsson G. Exposition and interpretation of the topology of neural networks // Proc. of the 18th IEEE International Conference on Machine Learning and Applications (ICMLA). 2019. P. 1069–1076.</mixed-citation><mixed-citation xml:lang="en">Gabrielsson R.B., Carlsson G. Exposition and interpretation of the topology of neural networks. Proc. of the 18th IEEE International Conference on Machine Learning and Applications (ICMLA), 2019, pp. 1069–1076.</mixed-citation></citation-alternatives></ref><ref id="cit15"><label>15</label><citation-alternatives><mixed-citation xml:lang="ru">Rieck B., Togninalli M., Bock C., Moor M., Horn M., Gumbsch T., Borwardt K. Neural persistence: A complexity measure for deep neural networks using algebraic topology // Proc. of the 7th International Conference on Learning Representations (ICLR). 2019.</mixed-citation><mixed-citation xml:lang="en">Rieck B., Togninalli M., Bock C., Moor M., Horn M., Gumbsch T., Borwardt K. Neural persistence: A complexity measure for deep neural networks using algebraic topology. Proc. of the 7th International Conference on Learning Representations (ICLR), 2019.</mixed-citation></citation-alternatives></ref><ref id="cit16"><label>16</label><citation-alternatives><mixed-citation xml:lang="ru">McGuire S., Jackson S., Emerson T., Kvinge H. Do neural networks trained with topological features learn different internal representations? // Proceedings of Machine Learning Research. 2023. V. 197. P. 122–136.</mixed-citation><mixed-citation xml:lang="en">McGuire S., Jackson S., Emerson T., Kvinge H. Do neural networks trained with topological features learn different internal representations? Proceedings of Machine Learning Research, 2023, vol. 197, pp. 122–136.</mixed-citation></citation-alternatives></ref><ref id="cit17"><label>17</label><citation-alternatives><mixed-citation xml:lang="ru">Guss W.H., Salakhutdinov R. On characterizing the capacity of neural networks using algebraic topology // arXiv. 2018. arXiv:1802.04443v1. https://doi.org/10.48550/arXiv.1802.04443</mixed-citation><mixed-citation xml:lang="en">Guss W.H., Salakhutdinov R. On characterizing the capacity of neural networks using algebraic topology. arXiv, 2018, arXiv:1802.04443v1. https://doi.org/10.48550/arXiv.1802.04443</mixed-citation></citation-alternatives></ref><ref id="cit18"><label>18</label><citation-alternatives><mixed-citation xml:lang="ru">Bergomi M.G., Frosini P., Giorgi D., Quercioli N. Towards a topological–geometrical theory of group equivariant non-expansive operators for data analysis and machine learning // Nature Machine Intelligence. 2019. V. 1. N 9. P. 423–433. https://doi.org/10.1038/s42256-019-0087-3</mixed-citation><mixed-citation xml:lang="en">Bergomi M.G., Frosini P., Giorgi D., Quercioli N. Towards a topological–geometrical theory of group equivariant non-expansive operators for data analysis and machine learning. Nature Machine Intelligence, 2019, vol. 1, no. 9, pp. 423–433. https://doi.org/10.1038/s42256-019-0087-3</mixed-citation></citation-alternatives></ref><ref id="cit19"><label>19</label><citation-alternatives><mixed-citation xml:lang="ru">Hofer C.D., Graf F., Niethammer M., Kwitt R. Topologically densified distributions // Proc. of the 37th International Conference on Machine Learning (ICML). 2020. P. 4254–4263.</mixed-citation><mixed-citation xml:lang="en">Hofer C.D., Graf F., Niethammer M., Kwitt R. Topologically densified distributions. Proc. of the 37th International Conference on Machine Learning (ICML), 2020, pp. 4254–4263.</mixed-citation></citation-alternatives></ref><ref id="cit20"><label>20</label><citation-alternatives><mixed-citation xml:lang="ru">Naitzat G., Zhitnikov A., Lim L.-H. Topology of deep neural networks // The Journal of Machine Learning Research. 2020. V. 21. N 1. P. 7503–7542.</mixed-citation><mixed-citation xml:lang="en">Naitzat G., Zhitnikov A., Lim L.-H. Topology of deep neural networks. The Journal of Machine Learning Research, 2020, vol. 21, no. 1, pp. 7503–7542.</mixed-citation></citation-alternatives></ref><ref id="cit21"><label>21</label><citation-alternatives><mixed-citation xml:lang="ru">Gal Y., Ghahramani Z. Dropout as a Bayesian approximation: Representing model uncertainty in deep learning // Proc. of the 33rd International Conference on Machine Learning (ICML). 2016. P. 1651–1660.</mixed-citation><mixed-citation xml:lang="en">Gal Y., Ghahramani Z. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. Proc. of the 33rd International Conference on Machine Learning (ICML), 2016, pp. 1651–1660.</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
