5 June 2023
Inaugural Lecture by Prof. Dr. Sören Laue, Informatics Dept., University of Hamburg
As machine learning continues to revolutionize various industries and everyday life, the need for efficient methods while maintaining interpretability becomes increasingly crucial. In this talk, I will explore the realm of efficient methods for machine learning, with a specific focus on white-box approaches that provide transparent and interpretable models.
One such white-box approach is to specify the machine learning problem as an optimization problem. Doing so provides control over various aspects of the problem, like different task objectives, noise models, or fairness constraints that must be met. I will present our approach GENO (GENeric Optimization), a framework for dealing with optimization problems originating from machine learning problems. GENO can automatically provide highly efficient solvers, and at the same time, the models are interpretable.
In conclusion, I emphasize the importance and advantages of white-box models and knowing the algorithmic details of a machine-learning method.
Further information about the speaker can be found here:
Date and time: 05.06.2023 17:15, Konrad-Zuse Hörsaal (B-201), Informatics Campus, University of Hamburg