Speaker: Francesco Verdoja
Robotics Seminar Series. Seventh Session – 24th June 2020, 13:00-14:00, via zoom. Link to event: https://aalto.zoom.us/j/62124942899
Among the various options to estimate uncertainty in deep neural networks, Monte-Carlo dropout is widely popular for its simplicity and effectiveness. However the quality of the uncertainty estimated through this method varies and choices in architecture design and in training procedures have to be carefully considered and tested to obtain satisfactory results. In this talk I will present a study offering a different point of view on the behavior of Monte-Carlo dropout, which will enable us to observe a few interesting properties of the technique to keep in mind when considering its use for uncertainty estimation.