Machine Learning
Contact : Olivier Fercoq
Machine Learning is at the intersection of many fields: applied mathematics, statistics, computer science and signal processing. One of its main objectives is to automatically learn to perform tasks based on observed examples.
Research in (machine) learning at Télécom Paris focuses on both theoretical aspects from statistics and optimization, as well as algorithmic aspects, to tackle the following issues:
- Variable selection, ranking, recommendation, structured prediction, anomaly detection, active learning or reinforcement learning…
The tools we develop are based on:
- Convex optimization, parsimonious optimization and regression, matrix factorization, kernel methods, stochastic approximation, statistical learning theory, extreme values …
Our research covers a wide range of applications:
- Neuroscience and brain imaging, audio/multimedia/image processing, pattern recognition, natural language processing, social networks, bioinformatics, predictive maintenance …
