Eun-Ah Kim, the Hans A. Bethe Professor of Physics in the College of Arts and Sciences at Cornell University. She has achieved outstanding breakthroughs in the field of quantum materials discovery. Leading a collaborative effort with Leslie Schoop, an associate professor of chemistry at Princeton University, Kim’s research focuses on integrating human intuition with artificial intelligence (AI) to enhance the search for new materials. Their efforts resulted in a first-of-its-kind analysis, released in Communications Materials. Its feature is a new machine-learning model that learns to replicate the valuable insights provided by human experts.
The innovative approach dubbed “Materials Expert-Artificial Intelligence” (ME-AI) specifically targets topological semimetals in square-net materials. Kim and her team used this model to test a dataset of 879 materials. Next, they ran a series of analyses to determine which of these had desirable characteristics. Kim pointed to the dangers of putting too much stock in AI. She underscored the importance of human judgment, arguing that there are certain physical characteristics that can only be fully grasped through expert intuition.
As part of her research, Kim articulated the importance of coupling AI with human intuition, especially in data-heavy pursuits like exploration. She stressed that emerging AI technologies are amazing, powerful tools. Yet, even these could not be trained to match the complex logic of human specialists. “Understanding certain materials requires a human expert’s reasoning and intuition,” said Kim, highlighting the importance of strategic integration of AI in material discovery.
Kim’s team worked hand in hand with Schoop’s research group. Collectively, they intentionally reviewed and tagged every piece of data they used to train their machine-learning model. This partnership is an excellent example of how interdisciplinary collaborations between industry and academia will continue to speed innovations in materials science.
Kim’s mission is to find next-generation materials through focused search as opposed to more serendipitous discoveries. The ME-AI model can significantly reduce the complexity of this task. It could be the catalyst needed to create innovations in everything from electronics to quantum computing.
Kim and her collaborators are leveraging the best of human expertise and AI power. As allies, partners, and sometimes competitors, they are paving the way for a new paradigm in materials discovery. Their research underscores the importance of expert intuition to scientific discovery. It demonstrates just how far AI can go and be successful without the incorporation of human expertise, or in tandem with it.