Division/Committee: [CINF: Division of Chemical Information] [COMP: Division of Computers in Chemistry]

Generative modeling is an emerging machine learning algorithm where complex objects, such as molecular structures, can be generated based on specific input requirements. From the initial examples of variational autoencoders and recurrent neural networks to the recent rise of large language models and diffusion models, generative modeling has been extensively applied in solving problems within the chemical sciences. These applications range from generating molecule or protein structures and predicting reaction outcomes to planning experiments. In this symposium, we welcome contributions that highlight the application of generative modeling across a broad spectrum of applications including but not limited to designing new molecular entities, enhancing conformation sampling, predicting transition states, and planning experimental procedures. Our goal is to showcase the diversity and similarities in methodologies applied to these issues, discuss areas for methodological innovation, and inspire new ideas for employing generative modeling in advancing molecular science.

Prof. Connor W. Coley

Massachusetts Institute of Technology

Appearances