[Colloquium] Learning Quantum Matter

Speaker:  Eun-Ah Kim (Cornell)

Title: Learning Quantum Matter

Abstract: The rapidly advancing intersection of quantum matter and artificial intelligence is redefining how we discover, characterize, and understand complex quantum phenomena. In this talk, I will outline our progress toward building a principled, data-driven framework for learning quantum matter. Beginning from materials informatics, we show how graphlet-based representations of crystalline materials enable interpretable machine learning to predict superconducting critical temperatures leveraging exhaustive collection of experimental data. I will then discuss X-TEC, an unsupervised framework for mapping phase evolution in large experimental datasets such as diffraction and microscopy, which has led to discoveries of novel phases. Moving from materials to quantum information, I will introduce the Quantum Attention Network (QuAN)—a physics-inspired neural architecture that learns to recognize complexity in many-qubit states—and a multi-core circuit decoder for quantum error correction. Finally, I will touch on emerging efforts to build AI assistants for theoretical research, where large language models can help automate routine calculations and code generation. Together, these efforts illustrate how physical insight, data science, and AI co-evolve toward a new scientific paradigm for understanding and designing quantum matter.