Jae Hyeon Lee

Title: Jae Hyeon Lee and Machine-Learning-Enabled Molecular Design.

Introduction

Jae Hyeon Lee is an inventor based in Boston, Massachusetts, and works at Genentech, Inc. His patent work concerns the use of machine learning to predict and design molecular structures and properties.

Latest Patents

The provided information identifies one patent associated with Lee: “Machine learning enabled prediction of molecular structures and properties.” The method receives a file describing a molecule’s initial three-dimensional structure and creates a representation of that molecule. This representation may use coarse-grained nodes corresponding to structural bodies of atoms in amino acid residues, or frames describing residue-backbone geometry and side-chain torsion angles. A design computation model then modifies the representation to determine a three-dimensional molecular structure associated with a desirable property or intended for a downstream task. Lee is credited with 1 patent.

Career Highlights

Lee’s listed work at Genentech, Inc. connects his inventive contribution with machine-learning approaches to molecular design. The patent describes computational methods for representing and modifying molecular structures, with potential relevance to designing molecules for specific properties or applications.

Collaborations

The provided information lists Payman Yadollahpour and Andrew Martin Watkins as Lee’s coworkers. Their inclusion indicates professional connections, but the available details do not specify their roles in the patent or whether they collaborated on it.

Conclusion

Jae Hyeon Lee’s identified patent applies machine learning to molecular structure prediction and design. It reflects an approach that combines molecular representations with computational modification to support desired properties and downstream uses.