Universidad de Guanajuato Applies Deep Learning to Predict Oxide Formation on Molybdenum Targets

Researchers at the Universidad de Guanajuato have developed a deep‑learning method employing convolutional neural networks to segment images and predict the diameter of oxide spots formed when thin molybdenum plates are irradiated with ultrashort laser pulses. The system processes visual data to identify different oxide types and accurately estimate spot dimensions based on exposure time and laser fluence. Results show that artificial‑intelligence techniques can automate and improve material‑characterisation tasks, offering a pathway for broader AI integration in materials science research.

The study, presented as a master’s thesis in Electrical Engineering (Instrumentation and Digital Systems), demonstrates the potential of AI‑driven analysis for extreme‑condition material behavior and may support future investigations of related processes.