course: 2AMU20

Machine learning, the scientific field which studies how machines (computers) can learn from data, can be roughly organized into discriminative learning and generative learning. Discriminative learning, the traditionally pre-dominant discipline in machine learning, is focusing on a teacher-student form of learning, i.e. on a learning scenario where there is a pre-specified form of questions to be asked, and a way to assess whether answers to these question are correct (or “how correct” they are). Generative learning, on the other hand, does not a-priori fix the specific type of questions we might want to answer with the model. Rather, in generative learning, data is provided “as is”, and the task of a generative model is to capture the nature of the data, the process generating it, and to discover interesting structure within the data.

Details

Level:

Master

Quartile:

Q4

Lecturers:

Cassio de Campos

Thomas Krak

Jakub Tomczak