Την ερχόμενη Παρασκευή 13/12 θα έχουμε τη χαρά να υποδεχτούμε ως ομιλητή στα πλαίσια της σειράς εκδηλώσεων του Τμήματος και των ΔΠΜΣ ΣΜΗΝ, ΥΔΑ και ΟΣΥΛ “Σεμινάριο CEID & Social Hour” τον κ. Βαγγέλη Μέτση, Αναπλ. Καθηγητή στο Texas State University. Θα ακολουθήσουν συζήτηση, καφές, κ.λπ. (εκτός κυλικείου!) Σημειώνεται ότι οι εκδηλώσεις υποστηρίζονται από το Τμήμα καθώς και από το ΔΠΜΣ ΣΜΗΝ. Ευχαριστούμε επίσης τη συνδρομή των φοιτητών του ACM Student Group. Επισημαίνουμε επίσης την σημαντική για όλους (και ιδιαίτερα τους φοιτητές μας) Ημερίδα Αποφοίτων που θα διεξαχθεί το Σάββατο 14/12, 11-2:30μμ. Το ακριβές πρόγραμμα θα ανακοινωθεί σύντομα.
Please note the following interesting talk that will be presented in the context of the weekly event “CEID Seminar & Social Hour” organized by CEID, and the MS programs DDCDM, SMIN and IHSS.
Title: From Signals to Insights: Advancing Time Series Analysis in Health and Biosignal Processing
Speaker: Dr. Vangelis Metsis, Associate Professor, Department of Computer Science, Texas State University.
Ημερομηνία-χώρος: Παρασκευή 6 Δεκεμβρίου, 3-5μμ, ΤΜΗΥΠ, αμφιθέατρο Γ
Abstract: This talk presents recent advancements in time series analysis, with a focus on health and biosignal processing applications. We begin by addressing fundamental challenges in time series data collection and labeling, introducing novel semi-automatic techniques to enhance data quality and combat label noise. The discussion then progresses to state-of-the-art approaches in time series representation and modeling, comparing traditional recurrent neural networks with modern attention-based mechanisms and transformers. We explore innovative methods for multimodal time series analysis, showcasing techniques for fusing learned representations from diverse biosignals. A significant portion of the talk is dedicated to generative models for time series, including GAN-based approaches and diffusion models, which address critical issues of data scarcity and class imbalance in medical datasets. Throughout the presentation, we demonstrate how these advanced techniques transform raw signals into meaningful insights, drawing from our research in sleep monitoring, emotion recognition, and biosignal synthesis. We also discuss best practices for evaluating time series models, ensuring robust performance in health-related applications. The talk concludes with an overview of emerging trends and open challenges in time series analysis for health and biosignal processing, providing a roadmap for future research in this rapidly evolving field.
About the speaker: Dr. Vangelis Metsis is an Associate Professor in the Department of Computer Science at Texas State University and the Director of the Intelligent Multimodal Computing and Sensing (IMICS) Lab. He joined the department in August 2014. Dr. Metsis earned his Ph.D. in Computer Science from the University of Texas at Arlington (UTA) in 2011, focusing on human-centered multimodal data analysis. He holds a B.S. with honors in Computer Science from the Department of Informatics at the Athens University of Economics and Business (AUEB) in Greece. Dr. Metsis’s research interests center on Machine Learning and Computer Vision, with applications in Smart Health, Pervasive Computing, Affective Computing, and Human-AR/VR Interaction. His work frequently involves time-series sensor data, as well as human physiological and behavioral data. He has secured numerous research grants as both Principal Investigator and Co-Principal Investigator, with funding from the NSF, U.S. Department of Education, and industry partners. Dr. Metsis has published extensively in peer-reviewed journals and conferences, with over 70 publications to his name.
