The hub on Machine Learning and Massive Data Analysis
was created in January 2012 as a result of the growing activity of CMLA in the context of predictive modeling and data-driven applications. It addresses decision problems involving real data with systematic confrontation to concrete applications.
Given the fast evolution in the field of theoretical machine learning, it had become clear that further innovation would mainly result from understanding the challenges related to real-life applications where additional constraints on the process of data collection and specific decision criteria occur.
The group focuses on exploring digital data from internet, industry, and simulation through three machine learning or statistical approaches:
- Scoring and ranking nonparametric methods for high dimensional data,
- Graph data mining, modeling, and inference,
- Active learning and sequential optimization.