Song Jeong In, an undergraduate student in the School of Materials Science & Engineering at our university, part of Professor Cho Ki Sub's research team (Artificial Intelligence Materials Design Laboratory), received the Outstanding Paper Award at the 2025 The Korean Society for heat treatment Fall Academic Conference.

Student Song Jeong In presented research on the topic ‘Development of a Microstructure Generation Model Using a Continuous Conditional Image Generation Model’. Existing continuous conditional GAN models suffered from mode collapse when the dimensionality of conditions—such as heat treatment temperature and time—increased, making it difficult to stably generate microstructures.

To address this, Student Song Jung In utilized a diffusion-based approach called CCDM (Continuous Conditional Diffusion Model) to build a model capable of stably generating microstructures even under continuous changes in heat treatment conditions. This model stably reproduced microstructure shapes without mode collapse even under multidimensional conditions like annealing temperature/time and austenitizing temperature/time. It demonstrated high generalization performance, with pattern differences due to condition changes resembling actual experimental trends.

Student Song Jeong In's research has laid the groundwork for reliably predicting microstructural changes even under complex heat treatment conditions. It is expected to contribute extensively to future AI-based process design and material development research.