Location : Angers, France
Background : The topic of this project is artificial intelligence applied to molecular computational chemistry. For the chemists, the ambition of this project is to radically change the approach, developing artificial intelligence and optimization methods in order to explore efficiently the highly combinatorial molecular space. The recent abundance of data is an incredible opportunity, but also an additional challenge and therefore an added value to this project: we will develop original and highly scalable methods.
Job Description : The postdoctoral researcher will start the machine learning development based on already available DFT calculations databases. Firstly, the goal of the predictive models is to generate for a new uncalculated molecule, precise approximations for different important results, saving hundreds hours of computation and making a broader exploration of the molecular space feasible. In addition to predictive models, generation of new molecules with constraints on one or more characteristics (such as electronic energies, the number of synthesis steps, etc.) will be investigated. We expect to study the integration of neural networks (objective functions) with neighborhood algorithms (molecular space exploration), but also emerging techniques like Generative Adversarial Networks (GAN).
Requirements : The candidate should therefore have experience in machine learning or data sciences. Additional computational chemistry experience will be appreciated. Fluency in Python language and solid knowledge of machine learning algorithms is furthermore mandatory.
Contacts :
– Benoit Da Mota (benoit.damota@univ-angers.fr)
– Thomas Cauchy (thomas.cauchy@univ-angers.fr)