Since 2012, Qarma has developed a general expertise in machine learning and modern artificial intelligence, from theory and algorithms to applications. The team is composed of  about 15-20 members with complementary skills covering all fields of machine learning, as illustrated by Qarma’s training activities.

Research works at Qarma are currently focussed on three major axes :

• learning theory;
• deep learning;
• connections between machine learning and signal processing;

and two transverse axes closely related to the above major axes :

• multi-view learning;
• machine learning for neurosciences.

Qarma’s research has also been related to spectral learning, grammatical inference, multi-task learning, active learning, budget learning, scalability, applications to gesture, audio, video, image, language.