BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
CALSCALE:GREGORIAN
PRODID:-//WordPress - MECv6.5.6//EN
X-ORIGINAL-URL:https://cmc.deusto.eus/
X-WR-CALNAME:cmc.deusto.eus
X-WR-CALDESC:DeustoCCM - Chair of Computational Mathematics at University of Deusto
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-PUBLISHED-TTL:PT1H
X-MS-OLK-FORCEINSPECTOROPEN:TRUE
BEGIN:VEVENT
CLASS:PUBLIC
UID:MEC-fe8946bd00dd1178502ce6befbb2a29a@cmc.deusto.eus
DTSTART:20260604T140000Z
DTEND:20260604T160000Z
DTSTAMP:20260601T155400Z
CREATED:20260601
LAST-MODIFIED:20260601
PRIORITY:5
TRANSP:OPAQUE
SUMMARY:IKTrace Seminar: When Equations meet Data: Inverse Problems and Hybrid-Cooperative Learning
DESCRIPTION:Date: Thu. June 04, 2026 (16:00H)\nEvent: IKTrace Seminar (IMAT-CCM)\nGeneral Colloquium / Research-Oriented\nThe IKTrace Seminar (IMAT-CCM), is a new initiative conceived as a forum for academic exchange, discussion, and collaboration within and beyond our mathematics research group.\nTitle: When Equations meet Data: Inverse Problems and Hybrid-Cooperative Learning\nSpeaker: Prof. Roberto Morales\nAffiliation: Postdoctoral Researccher. DeustoCCM – Chair of Computational Mathematics, University of Deusto\nAbstract.  Mathematical models are essential tools for describing, understanding, and predicting real-world phenomena. Many of these models are written in terms of differential equations, which express how physical, biological, or technological systems evolve in time and space. However, in many applications, the model is only partially known: some parameters, sources, boundary conditions, or internal mechanisms must be recovered from indirect and often noisy observations. This leads naturally to the field of inverse problems.\nIn this talk, we will introduce the basic ideas behind mathematical modelling with differential equations and discuss how data-driven methods, particularly neural networks, can be used as flexible tools for function approximation. We will then explain the difference between direct and inverse problems, highlighting why inverse problems are challenging and important in applications such as heat transfer, medical imaging, geophysics, and scientific machine learning.\nFinally, we will present the main idea behind Hybrid-Cooperative Learning (HYCO), a framework that combines physics-based models with data-driven models through a cooperative interaction. The goal is to show how equations and data can work together to recover hidden information, improve predictions, and build more reliable models for applied problems. The talk is intended as an accessible introduction for students interested in applied mathematics, and the interaction between modelling, computation, and machine learning.\n\nWHERE\nUniversity of Deusto, DeustoTech.\nRoom: Logistar (4th floor, ESIDE Building)\nUnibertsitate Etorb., 24, 48007 Bilbao, Bizkaia\n\nPoster\nThe seminar is intended primarily as an in-person event. However, for those who are unable to attend in person, it will also be possible to connect via this Meet link.\nThe seminar aims to strengthen mutual awareness of ongoing research lines, teaching interests, and interdisciplinary initiatives that will help identify shared interests, foster new synergies, and encourage future collaborations among colleagues from different areas of the Faculty of Engineering at University of Deusto.\n
URL:https://cmc.deusto.eus/events-calendar/iktrace-seminar-when-equations-meet-data-inverse-problems-and-hybrid-cooperative-learning/
ATTACH;FMTTYPE=image/png:https://cmc.deusto.eus/wp-content/uploads/IKtraceseminar_04jun2026_img.png
END:VEVENT
END:VCALENDAR