Ph.D Student
CurrentI am doing my Ph.D. on Improving Recent Self-Supervised Visual Representation Learning methods. More generally, My Research Aspires to Reduce Shortcut Learning, which Limits the Generalization and Robustness Capacity of Machine Learning, thus Restricting the Confidence we can put in such Algorithms.My First Contribution resulted in Two Papers (one is Published, and the other is still Under Revision) and proposed an Improvement to Invariance-based Instance Discrimination methods via an Equivariance task.I am now at the beginning of a Second Contribution which consists in studying how to Evaluate and Reduce Shortcut Learning in Self-Supervised Representation Learning approaches to find ways to Improve these Methods.