Exploring Transfer Learning techniques with Generative AI for the prediction of chemical processes performances

IFP Energies nouvelles (IFPEN) est un acteur majeur de la recherche et de la formation dans les domaines de l’énergie, du transport et de l’environnement. De la recherche à l’industrie, l’innovation technologique est au cœur de son action, articulée autour de quatre priorités stratégiques : Mobilité Durable, Energies Nouvelles, Climat / Environnement / Economie circulaire et Hydrocarbures Responsables.

Dans le cadre de la mission d’intérêt général confiée par les pouvoirs publics, IFPEN concentre ses efforts sur :

• l’apport de solutions aux défis sociétaux de l’énergie et du climat, en favorisant la transition vers une mobilité durable et l’émergence d’un mix énergétique plus diversifié ;
• la création de richesse et d’emplois, en soutenant l’activité économique française et européenne et la compétitivité des filières industrielles associées.

Partie intégrante d’IFPEN, l’école d’ingénieurs IFP School prépare les générations futures à relever ces défis.

Exploring Transfer Learning techniques with Generative AI for the prediction of chemical processes performances

IFPEN is an important player in the triple energy, ecological, and digital transition by offering differentiating technological solutions in response to societal and industrial challenges of energy and climate. The implementation of new methodological approaches combining "data science and experimentation" is among the studied solutions that allow for faster progress and reduced R&I costs.

The prediction of the output impurities content, such as sulphur or nitrogen, is a key factor when developing new catalysts or new processes. Data scarcity and poor generalization to new experimental conditions often limit the quality of the kinetic models or even and standard machine learning techniques.

One of the solutions for improving models is reusing knowledge from previous datasets. Transfer Learning is a promising approach to model new catalysts or processes. Previous studies conducted at IFPEN led to important improvements using a Bayesian approach. Other techniques, that use Generative Adversarial Networks (GANs), along with feature augmentation, allow model’s deep understanding of the dataset’s feature distribution, thus improving model training and robustness.

Internship objective

• Conduct an in-depth literature review of transfer learning with generative techniques (e.g. GANs), with a focus on their applications in chemical processes.
• Adapt and implement selected approaches.
• Test and evaluate the proposed methods using simulated datasets.

Profile

We are seeking a candidate with an engineering degree or pursuing a Master’s (M2) in Applied Mathematics, Artificial Intelligence or Data Science. Chemical engineering students with AI background are also encouraged to apply.

• Technical Skills: Strong foundation in deep learning and machine learning, with proficiency in Python and experience with deep learning frameworks.
• Knowledge Areas: Familiarity with Transfer Learning, generative models (e.g., GANs), and data augmentation techniques is highly desirable.
• Soft Skills: Analytical mindset, problem-solving abilities, a willingness to learn and adapt to new aspects, interests in AI.