Keeping pace with the rapid expansion of Artificial Intelligence and Machine Learning as major directions in contemporary research, the seminar titled “The Implement of Machine Learning and AI in Recent Research and Development” was successfully held on the afternoon of 07 March 2025 under the leadership of Assoc. Prof. Chi Phan, bringing significant academic value and meaningful networking opportunities.

Assoc. Prof. Chi Phan opened the seminar by introducing the objectives of the event and presenting an overview of VinUniversity as a young institution with strong ambitions in research excellence, international collaboration, and innovation-driven education. The seminar then featured presentations from speakers representing different research domains, each highlighting how AI and machine learning methods are being integrated into modern scientific practice.

The first invited keynote speaker was Professor Christian Soize, Emeritus Professor at Université Gustave Eiffel and a leading authority in uncertainty quantification for complex physical systems. In his presentation, Professor Soize discussed the fundamental challenges of parameter uncertainty and model uncertainty in computational models. He presented advanced probabilistic approaches for representing these uncertainties through tensor-valued random fields, random operators, and nonparametric probabilistic frameworks. These concepts are particularly important for modern AI systems, where training data may be limited and models must remain robust despite incomplete knowledge of physical processes. Professor Soize also discussed advanced techniques that enable reliable model learning even with small datasets, a challenge frequently encountered in scientific and engineering applications of AI.

Following this presentation, Dr. Tuan Nguyen-Sy, Head of the Laboratory for Computational Mechanics at Van Lang University, presented recent work on Generative AI for engineering design. His talk focused on the use of Conditional Variational Autoencoders (CVAE) to support automated HVAC system design in the Architecture Engineering Construction domain. By training generative models on millions of synthetic images, the approach enables the rapid generation of thousands of feasible design configurations within minutes. This significantly accelerates early-stage design exploration while still allowing engineers to evaluate and refine AI-generated solutions. The presentation highlighted how generative AI can act as a collaborative design tool that augments, rather than replaces, expert decision-making.

Another important theme of the seminar concerned the uncertainty of machine learning models themselves, a rapidly developing research area. Recent studies on machine learning uncertainty quantification demonstrate how theoretical frameworks originally developed by Professor Soize for stochastic mechanical systems can inform modern AI models. These developments illustrate how rigorous probabilistic theory can enhance the reliability and interpretability of machine learning systems used in scientific applications.

In the pharmaceutical field, Dr. Mai Dung Do, Head of the Computational and Artificial Intelligence Application Unit at Hanoi University of Pharmacy, discussed the transformative role of AI in drug discovery and development. She emphasized that AI techniques can significantly accelerate the identification of promising drug candidates and reduce the need for extensive experimental screening, thereby lowering the cost and time required for pharmaceutical innovation. Addressing the question “Can AI replace doctors?”, Dr. Mai Dung Do emphasized the collaborative role of AI in supporting doctors and medical experts in diagnosis, examination, and clinical decision-making rather than replacing human expertise. She also noted that AI-supported systems are increasingly being adopted in hospitals, bringing significant benefits to healthcare systems.

The seminar concluded with a presentation by Assoc. Prof. Quy Dong To from Université Gustave Eiffel, who introduced the structure of academic programs in France as well as graduate scholarship opportunities available for students wishing to pursue further studies there. Beyond the formal presentations, students had the opportunity to directly interact with the speakers through discussions and informal exchanges.

The visiting scholars also participated in campus visits and photo activities around VinUniversity, contributing to a dynamic and engaging academic atmosphere.