?
Topology and geometry of data manifold in deep learning
Cornell University
,
2022.
Despite significant advances in the field of deep learning in applications to various fields, explaining the inner processes of deep learning models remains an important and open question. The purpose of this article is to describe and substantiate the geometric and topological view of the learning process of neural networks. Our attention is focused on the internal representation of neural networks and on the dynamics of changes in the topology and geometry of the data manifold on different layers. We also propose a method for assessing the generalizing ability of neural networks based on topological descriptors. In this paper, we use the concepts of topological data analysis and intrinsic dimension, and we present a wide range of experiments on different datasets and different configurations of convolutional neural network architectures. In addition, we consider the issue of the geometry of adversarial attacks in the classification task and spoofing attacks on face recognition systems. Our work is a contribution to the development of an important area of explainable and interpretable AI through the example of computer vision.
Priority areas:
IT and mathematics
Language:
English
М. : Торус Пресс, 2018
The volume contains the abstracts of the 12th International Conference "Intelligent Data Processing: Theory and Applications". The conference is organized by the Russian Academy of Sciences, the Federal Research Center "Informatics and Control" of the Russian Academy of Sciences and the Scientific and Coordination Center "Digital Methods of Data Mining". The conference has being held biennially since 1989. It is one ...
Added: October 9, 2018
IEEE, 2020
Added: October 3, 2020
CEUR Workshop Proceedings, 2019
Added: October 31, 2019
Zhukova L., Кирюшина А. А., Ковальчук И. М. et al., Прикаспийский журнал: управление и высокие технологии 2018 № 2 (42) С. 56-68
The authors of the article conducted a study aimed at studying the problems associated with this form of training and the reasons for their occurrence. The article considers an example of the application of one of the machine learning methods – cluster analysis and describes the blockchain technology adapted for the implementation of the control ...
Added: October 18, 2018
Кириллов А. Н., Гавриков М. И., Lobacheva E. et al., Интеллектуальные системы. Теория и приложения 2015 Т. 19 № 2 С. 75-95
In this paper we consider the Shape Boltzmann Machine(SBM) and its multi-label version MSBM. We present an algorithm for training MSBM using only binary masks of objects and the seeds which approximately correspond to the locations of objects parts. ...
Added: September 30, 2015
CEUR-WS.org, 2020
The CLA conference is an international forum for researchers, practitioners and students dedicated to the practice of Formal Concept Analysis (FCA) and areas closely related to it, including data analysis and mining, information retrieval, knowledge management, knowledge engineering, logic, algebra and lattice theory.
The 15th of CLA, CLA 2020, was going to be held in Tallinn, Estonia ...
Added: October 30, 2020
Aachen : CEUR Workshop Proceedings, 2019
Workshop concentrates on an interdisciplinary approach to modelling human behavior incorporating data mining and expert knowledge from behavioral sciences. Data analysis results extracted from clean data of laboratory experiments will be compared with noisy industrial datasets from the web e.g. Insights from behavioral sciences will help data scientists. Behavior scientists will see new inspirations to ...
Added: November 19, 2019
Zhuk R., Ignatov D. I., Konstantinova N., Procedia Computer Science 2014 Vol. 31 P. 928-938
We propose extensions of the classical JSM-method and the Na ̈ıve Bayesian classifier for the case of triadic relational data. We performed a series of experiments on various types of data (both real and synthetic) to estimate quality of classification techniques and compare them with other classification algorithms that generate hypotheses, e.g. ID3 and Random ...
Added: June 9, 2014
University Rennes 1, 2017
This volume is the supplementary volume of the 14th International Conference on Formal Concept Analysis (ICFCA 2017), held from June 13th to 16th 2017, at IRISA, Rennes. The ICFCA conference series is one of the major venues for researches from the field of Formal Concept Analysis and related areas to present and discuss their recent ...
Added: June 19, 2017
Springer, 2021
This book constitutes the proceedings of the 16th International Conference on Formal Concept Analysis, ICFCA 2021, held in Strasbourg, France, in June/July 2021.
The 14 full papers and 5 short papers presented in this volume were carefully reviewed and selected from 32 submissions. The book also contains four invited contributions in full paper length.
The research part ...
Added: July 10, 2021
Kuznetsov V. O., Логистика и управление цепями поставок 2018 № 4 (87) С. 27-33
One of the options for a more flexible approach to analyzing the reliability of supply chains is the principal component analysis (PCA). With a large number of variables describing supply chain, it is a difficult task to analyze the structure of variables in two-dimensional space. Within the analysis of the variables dependencies PCA allows to ...
Added: November 29, 2018
Piscataway : IEEE, 2020
2020 International Joint Conference on Neural Networks (IJCNN) held virtually, as part of the IEEE World Congress on Computational Intelligence (IEEE WCCI) 2020. IJCNN 2020 is jointly organized by the IEEE Computational Intelligence Society (CIS) and the International Neural Network Society (INNS). For IJCNN 2020 (and when WCCI is organized in even-numbered years) IEEE CIS ...
Added: October 15, 2020
Sorokin K., Ayzenberg A., Анохин К. В. et al., / Cornell University. Series Computer Science "arxiv.org". 2022.
In present paper we test different approaches to reconstructing the topology of the physical space from neural activity data in A1 fields of mice brains, in particular, having a Cognitome-focused approach in mind. Animals were placed in different new environments and discovered them while their physical and neural activity was recorded. We discuss possible approaches ...
Added: November 17, 2022
V'yugin V., М. : МЦНМО, 2013
Книга предназначена для первоначлаьного знакомства с математическими основами современной теории машинного обучения (Machine Learning) и теории игр на предсказания. В первой части излагаются основы статистической теории машинного обучения, рассматриваются задачи классификации и регрессии с опорными векторами, теория обобщения и алгоритмы построения разделяющих гиперплоскостей. Во второй и третьей частях рассматриваются задачи адаптивного прогнозирования в нестохастических теоретико-игровой ...
Added: July 9, 2014
Нужный А. С., Однолько И. С., Глухов А. Ю. et al., Прикладная математика и вопросы управления 2021 № 1 С. 7-22
The paper proposes a mathematical model to optimize the operation of the tar hydrocracking unit.
The purpose of modeling is to improve the economic effect of product output by selecting optimal parameters,
such as hydrogen flow rate and reactor temperature. Hot Filtered Precipitation (HFT) is used as a target.
The model involves the search for the minimum value ...
Added: April 11, 2021
Ekaterinburg : CEUR Workshop Proceedings, 2014
AIST'2014 is an international data science conference on Analysis of Images, Social Networks, and Texts. Traditionally, the conference is held annually in Yekaterinburg, Russia. The conference is intended for computer scientists and practitioners whose research interests involve Internet mathematics and other related fields of data science.
LIST OF TOPICS (NON EXHAUSTIVE)
Applications of Data Mining and Machine ...
Added: August 28, 2014
Berlin : Association for Computational Linguistics, 2016
The 2016 Conference on Computational Natural Language Learning is the twentieth in the series of annual meetings organized by SIGNLL, the ACL special interest group on natural language learning. CoNLL 2016 will be held on August 11-12, 2016, and is co-located with the 54th annual meeting of the Association for Computational Linguistics (ACL) in Berlin, ...
Added: November 12, 2016
Switzerland : Springer, 2019
This book constitutes the refereed proceedings of the 11th International Conference on Intelligent Data Processing, IDP 2016, held in Barcelona, Spain, in October 2016.
The 11 revised full papers were carefully reviewed and selected from 52 submissions. The papers of this volume are organized in topical sections on machine learning theory with applications; intelligent data processing in life ...
Added: February 8, 2020
Krylova D., Maksimenko A., Государственное управление. Электронный вестник 2021 № 84 С. 241-255
In this article, the authors, using the example of several foreign publications, analyze the trends in the use of artificial intelligence and machine learning in discernment of corruption. Based on the international review, the authors make the conclusion that the mechanisms for detecting corruption, based on the use of artificial intelligence, described in foreign sources, ...
Added: February 25, 2021
Emmanuel I. C., Mitrofanova E., / Cornell Tech. Series 4064475 "ArXiv Preprint". 2022.
The paper is devoted to the study of the model fairness and process fairness of the Russian demographic dataset by making predictions of divorce of the 1st marriage, religiosity, 1st employment and completion of education. Our goal was to make classifiers more equitable by reducing their reliance on sensitive features while increasing or at least ...
Added: May 31, 2022
Bartunov S., Кондрашкин Д. А., Osokin A. et al., / Arxiv.org. Series arXiv:1502.07257 "Computation and language". 2015.
Recently proposed Skip-gram model is a powerful method for learning high-dimensional word representations that capture rich semantic relationships between words. However, Skip-gram as well as most prior work on learning word representations does not take into account word ambiguity and maintain only single representation per word. Although a number of Skip-gram modifications were proposed to ...
Added: November 5, 2015
Денис Турдаков, Астраханцев Н. А., Недумов Я. Р. et al., Труды Института системного программирования РАН 2014 Т. 26 С. 421-438
he paper presents a framework for fast text analytics developed during the Texterra project. Texterra is a technology for multilingual text mining based on novel text processing methods that exploit knowledge extracted from user-generated content. It delivers a fast scalable solution for text mining without the expensive customization. Depending on use-cases Texterra could be utilized ...
Added: November 6, 2017