Fakultät für Mathematik und Naturwissenschaften

MS05: Mathematics for Big Data

The talks in this minisymposium are addressing different mathematical challenges coming from Big Data problems, in broad sense - either in the sense of development of optimization and statistical methods for modelling and solving problems that arise in real life decision process, applications to modelling complex system and analysing their outputs.

The topics which will be covered in the minisymposium vary from optimization methods designed for machine learning problems, noisy optimization, adversarial attacks on neural networks, uncertainty quantification in deep learning, functional statistics and stochastic geometric models for face recognition and linear algebra methods for big data analysis.
Almost all topics are either coming from a specific industrial problem or are highly applicable in real life data analysis.
In particular, 6 talks are on industrial projects developed within H2020 MSC BIGMATH project (grant n. 812912).

Tuesday, April 13, 14:00-15:40

  1. 14:00 Stevo Rackovic, Claudia Soares, Dusan Jakovetic, Zoranka Desnica, Clustering of the Blendshape Facial Model
  2. 14:25 Rongjiao Ji, Alessandra Micheletti, Natasa Krklec Jerinkic, Zoranka Desnica, Emotion pattern detection on facial videos using functional statistics
  3. 14:50 Filipa Valdeira, Ricardo Ferreira, Alessandra Micheletti, Cláudia Soares, From noisy point clouds to complete ear models: unsupervised pipeline for application in the prosthetic industry
  4. 15:15 Giulia Ferrandi, Causality in a sequence of events via graph clustering

Tuesday, April 13, 16:00-17:40

  1. 16:00 Perfect Yayra Gidisu, Michiel Hochstenbach, Deterministic Subset Selection Algorithms
  2. 16:25 Greta Malaspina, Natasa Krejic, Lense Swaenen, A Modified Levenberg-Marquardt Method for Large Scale Network Adjustment
  3. 16:50 Nataša Krejić, Nataša Krklec Jerinkić, Tijana Ostojić, BFGS Method for Minimizing Nonsmooth Convex Functions with Variable Accuracy
  4. 17:15 Natasa Krejic, Milos Savic, Jasna Atanasijevic, Dusan Jakovetic, The HUNOD Method for Tax Evasion Risk Management

Wednesday, April 14, 14:00-15:40

  1. 14:00 Hanno Gottschalk, Hayk Asatryan, Marieke Lippert, Matthias Rottmann, Generative Adversarial Learning in Hölder Spaces
  2. 14:25 Matthias Rottmann, Kira Maag, Mathis Peyron, Nataša Krejić, Hanno Gottschalk, Detection of Iterative Adversarial Attacks on Deep Neural Networks via Counter Attack
  3. 14:50 Dusan Milisav Jakovetic, Anomaly Detection for Cybersecurity in Industrial Internet of Things
  4. 15:15 Jose Mario Martinez, Estimation of parameters in one-dimensional models of natural channels

Wednesday, April 14, 16:00-17:40

  1. 16:00 Stefania Bellavia, Jacek Gondzio, Margherita Porcelli, Relaxed interior Point methods for low-rank semidefinite programming problems with applications to Matrix Completion
  2. 16:25 Daniela di Serafino, Germana Landi, Marco Viola, Restoring Poissonian Images with One-Directional Texture
  3. 16:50 Miguel David Bustamante, Optimal Scheduling of Distributed Generation to Achieve Linear Aggregate Response
  4. 17:15 Om Bishnu Chaudhary, Applications of Convolutional Neural Network

Check the SIG Webpage.

Check the BIGMATH Webpage.

Weitere Infos über #UniWuppertal: