It is common knowledge that numerous algorithms deployed for artificial intelligence (AI) are based on mathematical principles. While present day AI research focuses on the application of AI over a wide array of fields such as image recognition, weather forecasts and disease diagnostics, the key to better harnessing the power of AI and upgrading its performance lies in better understanding and studying mathematics as the underlying discipline that enables AI.

Prior to 2016, the Applied Mathematics and Mining Laboratory led by professor Hyung Ju Hwang at the Department of Mathematics, POSTECH, mainly concentrated on nonlinear partial differential equations and convergence among physics, machinery and bioscience. Since then, the Lab has shifted its focus to ‘artificial intelligence’ driven by mathematics – such as capitalizing on neural networks to solve differential equations. This has enabled the Lab to fully dedicate itself to developing math-AI hybrid models and deep learning-powered mathematical methodologies.

The shift certainly freed up the Lab to extend its horizons – from exploring fundamental mathematical theories to undertaking various projects with businesses and organizations. In 2017, the Lab worked with a pharmaceutical company to conduct predictive analyses on cancer patients, and then with Samsung to develop a methodology for detecting anomalies in industrial data based on artificial intelligence with Samsung Electronics. Besides mentioned above, the Lab also delved into Corporate bankruptcy prediction and investment algorithms, generating noteworthy outcomes in resolving challenges on the shop floor and for society at large across the manufacturing, healthcare and fintech.

In particular, research findings, generated through the application of AI to predict the endpoint temperatures of converters used to remove carbon from the molten iron coming from the blast furnace, were adopted by POSCO to make significant contributions to mitigating costs on the shop floor. It also enabled the Lab to earn the ‘Open & Collaboration Award’ at the POSCO Technology Conference 2019. Recently, the Lab analyzed the efficiency of the Korean government’s disease control policy by leveraging COVID-19 data and AI, and published its outcomes in a prominent medical informatics journal.

The Applied Mathematics and Mining Laboratory firmly believes that ‘it is the most fundamental theories that translate into the best applications’, and recognizes the value of an exploratory mindset that ‘never fears failure but embraces it as a step towards creating something entirely new’, making valuable contributions to the resolution of industrial and social challenges, and thereby, helping build a better tomorrow for everyone.