Deep Learning in Synthesis Planning
Organic synthesis is a critical discipline that directly brings scientific and societal benefits by enabling access to high added-value molecules and to new molecules. Among all the methods for organic synthesis, retrosynthetic analysis is a problem-solving technique widely used by organic chemists. Computational retrosynthetic analysis tools can greatly assist this process and would have many applications in drug discovery, medicinal chemistry, material science, etc. However, most computer-aided retrosynthetic tools are slow and provide results of unsatisfactory quality. Therefore, researchers are introducing machine learning into this field to fix these problems. In this project, we will employ deep neural network to solve the challenging problem of computer-aided retrosynthetic analysis, discover new efficient synthetic routes for organic molecules, and bring new insights into the cross-domain of computer science and chemistry.
Using deep learning algorithm to perform retrosynthesis for organic molecule.
Applicant's Learning Objectives
training in machine learning methods; research experiences in organic retrosynthesis
Complexity of the project