Research area

The team Machine Learning and Applications (A3) tackles machine learning problems by using supervised, unsupervised and hybrid methods. This research is fed, coordinated and evaluated thanks to various applications in the field of the pattern recognition and data mining.

Topic 1: Theoretical and Algorithmic supervised learning

• Models for assessing relational learning
• Learning from ambiguous information and information theory

Topic 2: Learning by and for action

• Relational Reinforcement Learning
• Collective learning

Topic 3: Collaborative and incremental unsupervised learning

• Vertical, horizontal and directional collaboration
• Memory based Learning, Transfer learning, Learning from data stream, ...
• Unsupervised variables selection
• Structural and adaptive Under-sampling for unbalanced distributions

Topic 4: Learning Mixture Models

• Probabilistic approaches, geometric, hierarchical, hybrid, ...
• Hierarchical topological tree Model for data visualization
• Learning topological structure of HMM

Topic 5: Exploratory Analysis of Complex Data

• Mining and analysis of dynamics real networks and biological
• Classification of structured data in sequences and trees
• Galois lattices and abstraction

Last modified: Tuesday 17 January 2012

Contact for this webpage: Sebastien.Guerif at lipn.univ-paris13.fr