BIMSA >
ICMRA Seminar Series
ICMRA Seminar Series
From Variational Inequality Problem to Machine Learning Problem
From Variational Inequality Problem to Machine Learning Problem
Organizers
Speaker
Jeremiah Nkwegu Ezeora
Time
Monday, June 22, 2026 10:00 AM - 11:00 AM
Venue
A3-4-101
Online
Zoom 204 323 0165
(BIMSA)
Abstract
Fixed point problems and variational inequalities are cornerstones of nonlinear analysis, while machine learning optimization drives modern AI, their deep connections are often downplayed. This talk aims to unveil the profound mathematical equivalence between Fixed Point Problems, Variational Inequalities, and Machine Learning optimization. Through classical derivations and concrete algorithmic examples, including Gradient Descent, and Proximal Gradient methods, we demonstrate that training modern AI models is fundamentally a fixed-point iteration solving a variational inequality problem. This unified perspective not only clarifies the theoretical advantages of popular algorithms but also provides a good way of designing novel optimization methods(AI-models) with guaranteed convergence and stability.
Speaker Intro
Dr Jeremiah Nkwegu Ezeora is a Professor of Mathematics(Functional Analysis) at the University of Port Hartcourt, Nigeria. He graduated from the African University of Science and Technology, Abuja, Nigeria with a PhD degree in November, 2013 after obtaining a Postgraduate Diploma in Mathematics of the ICTP, Trieste, Italy. He has over 50 papers published in both locally and internationally highly rated journals. He has supervised 3 PhD students, 5 MSc students and mentored over 100 undergraduate students, many of them, now have PhD from different Universities globally. His research interests are in Fixed Point Theory, Differential Equations, Control Theory and Optimization Problems.
He is visiting BIMSA with support from the ICMRA Visiting Scholars Program.
He is visiting BIMSA with support from the ICMRA Visiting Scholars Program.