Abstract: In ML, much of the focus is on validating the model by, for example, measuring its accuracy and precision. However, traditional software has taught us that problems often lie in the integration of the components. This combined with the often whimsical nature of ML calls for methods to validate the systems utilizing ML as whole. In this talk, I will present methods that have been used in research literature to validate AI systems.

Speakers:  Lalli Myllyaho
Lalli is a PhD student at the University of Helsinki, working for IVVES project (https://ivves.eu/) in Empirical Software Engineering group.

Affiliation: University of Helsinki

Place of Seminar:  Zoom

Meeting ID: 653 3878 7437

Passcode: 838326