Invited Speakers of ICOMS 2025
Dr. Gökhan Yıldırım
Bilkent University, Turkey
Dr. Gökhan Yıldırım is a faculty member at Bilkent University, Ankara, Türkiye. He earned his PhD in Applied Mathematics from the University of Southern California, USA. His research spans enumerative combinatorics, probability, statistical physics, and data science. Dr. Yıldırım has contributed to numerous collaborative research projects and published in prestigious peer-reviewed international journals, including Advances in Applied Mathematics, Nature Physics, Electronic Journal of Probability, and Electronic Journal of Combinatorics. His findings have been presented at leading international conferences and meetings worldwide. In addition to his research, Dr. Yıldırım is an experienced educator, regularly teaching courses in probability, statistics, discrete mathematics, and data science. He has also served on editorial and review boards of scientific journals and actively participated in program committees of international conferences.
Speech Title: "Sampling from Discrete Combinatorial Structures: Algorithms, Challenges, and Open Questions"
Abstract: Understanding the structure and properties of discrete combinatorial objects, such as pattern-avoiding permutations and inversion sequences, is a fundamental problem with applications across mathematics, statistical physics, and computer science. In this talk, I will explore sampling algorithms designed to efficiently generate these complex structures. Special emphasis will be placed on Markov Chain Monte Carlo (MCMC) methods, which have become indispensable tools for approximate sampling in high-dimensional combinatorial spaces. I will discuss the strengths and limitations of these algorithms, highlighting recent advancements and illustrating their practical performance. Additionally, several open questions will be presented, offering insight into ongoing challenges in the field and potential directions for future research. This talk aims to provide an accessible overview of the theoretical underpinnings, computational techniques, and open problems related to sampling from discrete combinatorial structures.
Dr Chris G. Antonopoulos
University of Essex, UK
Dr Antonopoulos earned his BSc in Mathematics from the University of Crete, Greece, in 1999. In the same year, he joined the Department of Mathematics at the University of Patras, Greece, where he completed an MSc in Applied Mathematics in 2002 and a PhD in 2007. His doctoral research focused on the analytical and numerical study of stability and chaos in multi-dimensional Hamiltonian systems, investigating their transition from classical to statistical mechanics. Dr Antonopoulos is the author of 58 peer-reviewed journal articles, 3 editorials, and 9 publications in conference proceedings and book chapters. He has delivered 58 talks at international conferences, universities, and academic institutes, including 22 invited lectures. His work has received significant recognition, with an h-index of 24 and 2,776 citations (Google Scholar). Dr Antonopoulos has successfully supervised 1 MSc and 3 PhD students and is currently supervising 2 PhD candidates. He is a member of the London Mathematical Society (LMS) and the Institute of Mathematics and its Applications (IMA). Since July 2020, he has served as the IMA Representative for the School of Mathematics, Statistics and Actuarial Science (SMSAS) at the University of Essex.
Speech Title: "Network Inference Combining Mutual Information Rate and Statistical Tests"
Abstract: In this talk, we introduce a novel method that integrates information-theoretical and statistical techniques to infer connectivity in complex networks using time-series data. Grounded in Shannon entropy, the approach estimates the Mutual Information Rate between pairs of time-series and applies statistical significance tests using the false discovery rate method for multiple hypothesis testing to determine connectivity. We present the mathematical framework and demonstrate the method’s efficacy through applications to various systems, including correlated normal variates, coupled circle and logistic maps, coupled Lorenz systems, and coupled stochastic Kuramoto phase oscillators. We further explore the impact of noise on the methodology for networks of coupled stochastic Kuramoto oscillators and the influence of coupling heterogeneity on networks of coupled circle maps. Using receiver operating characteristic (ROC) curves, we show that the method accurately identifies connected nodes and pairs. In the context of stochastic data, it effectively reconstructs the original connectivity structure. Moreover, the methodology is shown to recover connectivity matrices for dynamics on Erdős-Rényi and small-world networks with diverse coupling heterogeneity. A key strength of this approach is its ability to infer the underlying network connectivity solely from recorded datasets, making it applicable to a wide range of fields, including functional network inference in neuroscience, financial market analysis, and social media networks. This versatility underscores its potential for analysing any network derived from time-series data.
Assist. Prof. Paola Lecca
Free University of Bozen-Bolzano, Italy
Paola Lecca is a theoretical physicist with a PhD in Computer Science and Telecommunication. Dr. Lecca is Assistant Professor at the Faculty of Engineering of the Free University of Bozen-Bolzano (Italy), where she carries out research activities in the fields of graph theory, modelling and analysis of dynamic networks, statistical inference, parallel computing, mainly in the application domains of bioinformatics, computational biology, biophysics and drug development. Dr. Lecca’s interests and scientific contributions relate to the mathematical foundations of data analysis, artificial intelligence and modelling techniques for complex systems that can be represented as networks of dynamic interactions. A member of various national and international scientific associations, Paola Lecca has also authored over a hundred articles in international scientific journals and conference proceedings in computational biology, biophysics and applied mathematics.
Speech Title: "The Bayesian basis of Decision-Making in Artificial Intelligence: Does It Help to Determine the Level of Confidence in Artificial Intelligence Results?"
Abstract: Artificial Intelligence systems can explicitly model uncertainty thanks to Bayesian statistics. Bayesian models produce a probability distribution over potential outcomes rather than a single deterministic output. By capturing the inherent uncertainty in data and model projections, this probabilistic approach facilitates risk assessment and well-informed decision-making. Furthermore, Bayesian decision theory provides a structured framework for making optimal decisions under uncertainty. However, the uncertainty in the prior probability function and background knowledge cannot be handled or represented generally by the Bayesian approach. This is a significant theoretical and practical drawback of Bayesianism. These limitations of the Bayesian approach will be presented and discussed in the talk and some perspectives will be given on the need to go beyond traditional Bayesianism.
Assoc. Prof. Ahmad Lutfi
Amri Ramli
Universiti
Sains Malaysia, Malaysia
Assoc. Prof. Dr. Ahmad Lutfi Amri Ramli is an esteemed academic at the School of Mathematical Sciences, Universiti Sains Malaysia (USM). He holds a PhD from Durham University, an MSc from Brunel University, and a BSc in Mathematics from USM. Dr. Lutfi’s research expertise lies in applied mathematics, with a focus on motion planning for robots integrating the knowledge from computer aided geometrical design, curve and surface modelling, and computational methods. His interest also extends to gamification in education, leveraging innovative approaches to enhance student engagement in mathematical learning As an academic leader and researcher, Dr. Lutfi has been actively involved in advancing research, fostering student development, and innovating teaching practices, contributing to the growth of mathematics and applied sciences.
Speech Title: "CAGD and Its Application in Transportation Research"
CAGD is used in transportation research for road design, vehicle trajectory planning, speed estimation, and optimizing autonomous navigation paths. It also enables the creation of smooth and efficient road alignments while ensuring safety and comfort. This talk will highlight a few key applications, such as safe speed estimation using parametric curves derived from GPS data and travel time prediction based on circular arc curvature to estimate speed from road geometry. These techniques provide valuable insights for traffic management and infrastructure development. By applying CAGD techniques, transportation research can leverage mathematical models to improve accuracy while reducing costs.
Prof. Eugene B. Postnikov
Kursk State University, Russia
Prof. Eugene B. Postnikov earned his PhD in Physics&Mathematics in 2000 and D.Sc. in 2011, and recently holds the position of the Head of the Theoretical Physics Department at the Research Center for Condensed Matter Physics and full professor at the Department of Physics and Nanotechnology, Kursk State University, Russia. His research interests relate to the mathematical modelling of dynamical and stochastic processes in complex media including thermodynamic and transport processes in liquids, soft matter and biophysical systems. He is an Editor in the journal ‘Chaos, Solitons & Fractals’ and an Editorial Board member of the "Control Theory and Mechanics" section in the journal 'Mathematics'.
Speech Title: "Padé Approximants to Data Exhibiting Fast Growth and Slow Saturation: From Superparamagnetic System to Water Swelling"
Abstract: A wide variety of complex microscopically heterogeneous media under an external influence exhibits response curves, which are hardly fitted and simulated by conventional simple kinetic models. Typical representatives of such systems are magnetization curves for concentrated magnetic fluids and soft matter responding to the external magnetic field and soft porous media (gels, sponges, etc.) swelling when placed in water or other liquids. Notable, such different physical systems can be treated mathematically in a similar way based on the low-order Padé approximation of experimental data, which exactly reproduces both asymptotic regimes of the process. It will be considered how the sequential approximation procedure can correspond to systems of kinetic equations, which reflect the principal features of a sequence of partially saturated capacities. Finally, more complex dynamic and structural phenomena, which may lead to fractional-order models, will be discussed.