COP 28


PhD on Machine Learning algorithm to predict Time Series

Léna Sasal

PhD student, Sorbonne Centre for Artificial Intelligence, SUAD

The prediction of time series is how to know what is going to happen tomorrow or next year based on what happened before (Yesterday, the month before, …). We are currently working on transformers which is one type of algorithm to perform such task and we can apply it on various field (ex: what will be the temperature tomorrow based on the last 2 years, what is the probability to have earthquake tomorrow, …). 

As we are working with Total Energies, our work consists in predicting how much oil will we produce tomorrow. The work done on this subject has led to the publication of an article (W-Transformers : A Wavelet-based Transformer Framework for Univariate Time Series Forecasting) that was presented at The International Conference on Machine Learning and Application in the Bahamas in December 2022.

This project is under the Chair of TotalEnergies, Sorbonne Centre for Artificial Intelligence, SUAD