GIST announced on the 6th that a research team led by Professor Hyunju Lee of the Department of AI developed an AI model called 'AlphaDTA' that estimates the performance of new drug candidates by leveraging internal information generated by AlphaFold3, Google DeepMind's protein structure prediction model. The research was published online on July 21 in the international journal 'Journal of Cheminformatics'. AlphaDTA analyzes both the 3D protein–drug structures predicted by AlphaFold3 and the internal representation information generated during structure computation. It showed higher predictive performance than existing sequence-based methods for new protein–drug combinations not used during training and produced results on par with existing structure-based methods even without experimental structural data. The team expects the technology to help rapidly screen new drug candidates lacking structural data and support drug repurposing research. Separately, UNIST announced on the 6th that a team led by Professor Yeon-Chang Lee of the Graduate School of Artificial Intelligence developed a cross-domain recommendation technology called 'Multi-TAP' that reads users' purchasing preferences by subdividing them into detailed product groups. The research was accepted for presentation at the ACM Knowledge Discovery and Data Mining (KDD) Conference, held for five days in Jeju Island starting August 9. Multi-TAP subdivides purchasing tendencies by detailed product group, has a large language model summarize them in sentences, and converts them into numerical information for the recommendation algorithm.