Multimodal Artificial Intelligence

The Multimodal Artificial Intelligence Research Group was established in September 2024, led by Fanjie Xu. Our team comprises members from diverse interdisciplinary backgrounds, including chemistry, biology, computer science, and artificial intelligence. We are dedicated to exploring multimodal learning and generative models for innovative scientific applications, with a particular focus on spectroscopy, materials science, and battery technology. By leveraging AI, we aim to uncover deep correlations within complex datasets, driving advancements in intelligent scientific research.

Research Directions

• Spectral-Structure Relationship
  Integrating spectroscopic data such as NMR, Raman, and XRD, we employ deep learning approaches to establish mappings between material structures and their spectral responses.
• Battery Design
  Utilizing generative models and multimodal learning, we optimize the structural and performance predictions of battery materials, accelerating the discovery and design of high-performance energy storage materials.
• Structural Analysis
  By incorporating experimental data from X-ray diffraction and electron microscopy, we leverage AI-driven methods for material structure analysis and crystal structure prediction.
• Molecular Representation
  Investigating molecular graph learning and multimodal transformers, we develop efficient representations for molecules and materials to enhance AI generalization in chemistry and materials science.

Research Achievements

Since its inception, the group has made significant progress in AI for Science and has published key research findings in leading international journals (including collaborative projects), with notable publications in Nat. Comput. Sci. and Adv. Sci..

Future Outlook

Moving forward, we will continue to integrate multimodal artificial intelligence with scientific computing, fostering data-driven scientific discovery. We welcome collaborations with researchers from both academia and industry to further explore the potential of AI in scientific research.