Post-doctoral position: Hydrogen storage materials designed by Machine Learning, Paris
143 views
Skip to first unread message
Nataliya Sokolovska
Jan 23, 2018, 3:25:29 PM1/23/18
to Machine Learning News
Postdoctoral position on “Hydrogen storage materials designed by Machine Learning”
1 year, to be started before 1st June 2018 in ICMPE – CNRS (Thiais, 8km south of Paris, France)
Objectives of the project
The storage of a "clean" energy is a key issue for the future, where hydrogen may play an important role. However, its storage still remains a technological challenge, where a safe and effective answer could be the use of metal hydrides, in which hydrogen atom is bonded to a metallic matrix. In this project, we are planning to use the machine learning (ML) approach to design new materials as metal hydrides for efficient hydrogen storage.
Job description
First of all, the postdoctoral researcher will build a robust database. This database will contain both crystallographic data, like pair distances distribution and stability/metastability information for each compound, like the heat of formation. The input data is coming from known crystal structures database and massive Density Functional Theory (DFT) calculations already partly performed.
Then, the postdoctoral researcher will evaluate several ML algorithms on different criteria like supervised or not, discriminative or generative. Among algorithms, several ways will be investigated as classical neural networks (deep learning), but also new and emerging techniques will be considered, like the Generative Adversarial Networks (GAN) which is already under investigation at the ICMPE. Finally, we expect to get prediction of new phases from an original chemical composition. In addition, a self-consistent procedure is envisaged: the stability of predicted new phases will be estimated by DFT and will be used as new input data using reinforcement learning.
Desired skills and experience
Eligible candidates should have a Ph.D. degree and experience in machine learning or data sciences, additional computational chemistry experience will be appreciated. The successful candidate should have substantial previous experiences in database building. Experience in scripting languages such as Matlab and/or Python is highly mandatory. A deep knowledge of machine learning algorithms is also fundamental. Daily interactions with solid state chemists will be required, and therefore a crystallography background is valuable but not mandatory.
About the employer
The ICMPE (Institut de Chimie et des Matériaux de Paris Est) is a CNRS/UPEC institute doing research on solid state chemistry. The team where the postdoc will be hires is skilled on crystallography, thermodynamic modelling and DFT calculations on metal hydrides. Machine learning is an emerging discipline in the team who collaborates already with partners in Paris.
Contacts
Application (detailed CV + motivation letter) and requests for more information has to be sent to Jean-Claude Crivello (criv...@icmpe.cnrs.fr) by email with “Apply for ML postdoc” as a subject.
1 year, to be started before 1st June 2018 in ICMPE – CNRS (Thiais, 8km south of Paris, France)
Objectives of the project
The storage of a "clean" energy is a key issue for the future, where hydrogen may play an important role. However, its storage still remains a technological challenge, where a safe and effective answer could be the use of metal hydrides, in which hydrogen atom is bonded to a metallic matrix. In this project, we are planning to use the machine learning (ML) approach to design new materials as metal hydrides for efficient hydrogen storage.
Job description
First of all, the postdoctoral researcher will build a robust database. This database will contain both crystallographic data, like pair distances distribution and stability/metastability information for each compound, like the heat of formation. The input data is coming from known crystal structures database and massive Density Functional Theory (DFT) calculations already partly performed.
Then, the postdoctoral researcher will evaluate several ML algorithms on different criteria like supervised or not, discriminative or generative. Among algorithms, several ways will be investigated as classical neural networks (deep learning), but also new and emerging techniques will be considered, like the Generative Adversarial Networks (GAN) which is already under investigation at the ICMPE. Finally, we expect to get prediction of new phases from an original chemical composition. In addition, a self-consistent procedure is envisaged: the stability of predicted new phases will be estimated by DFT and will be used as new input data using reinforcement learning.
Desired skills and experience
Eligible candidates should have a Ph.D. degree and experience in machine learning or data sciences, additional computational chemistry experience will be appreciated. The successful candidate should have substantial previous experiences in database building. Experience in scripting languages such as Matlab and/or Python is highly mandatory. A deep knowledge of machine learning algorithms is also fundamental. Daily interactions with solid state chemists will be required, and therefore a crystallography background is valuable but not mandatory.
About the employer
The ICMPE (Institut de Chimie et des Matériaux de Paris Est) is a CNRS/UPEC institute doing research on solid state chemistry. The team where the postdoc will be hires is skilled on crystallography, thermodynamic modelling and DFT calculations on metal hydrides. Machine learning is an emerging discipline in the team who collaborates already with partners in Paris.
Contacts
Application (detailed CV + motivation letter) and requests for more information has to be sent to Jean-Claude Crivello (criv...@icmpe.cnrs.fr) by email with “Apply for ML postdoc” as a subject.
Reply all
Reply to author
Forward
0 new messages