Junior Research Scientist in machine learning and deep learning on structured data

The Mathématiques et Informatique appliquées de Toulouse (MIAT) research unit is part of the Mathematics, computer and data sciences, digital technologies (MathNum) division of INRAE. The research unit is composed of two research teams (SaAB and SCIDyn) and three platforms (GENOTOUL Bioinfo, RECORD, and SIGENAE). In this laboratory, you will join the Statistics and Algorithmics for Biology (SaAB) team made up of ten researchers and engineers, a multi-disciplinary team in computer science/bioinformatics/statistics whose scientific applications are oriented towards biology. The team has developed skills on the topic of deep neural networks, and more particularly on neural networks capable of using or predicting structured information (for example graphs). The team also holds the "Design using intuition and logic" (DIL) chair at the Artificial and Natural Intelligence Toulouse Institute (ANITI).
You will strengthen the team's synergy around the central theme of machine learning and complement the skills already present on the theme of deep learning, As a specialist in this field, you wil be capable of proposing original solutions/architectures for learning structured objects with neural approaches.
One of the main methodological obstacles concerns the development of original hybridizations between discrete optimization (for structure learning) and continuous optimization (on which neural networks are based). These questions are of great interest for various problems addressed by the team: inference of gene networks, protein design, phenotype prediction, and automatic annotation of genomes.