As usual, we had researchers working on very different applications. From industrial quality control of machine parts and fruit sorting to cancer diagnostics with hyperspectral imaging, ECG brain signal monitoring, classification of mosquito species, gait recognition and text mining.

This group was very knowledgeable. Most participants had prior experience in using machine learning. Those who did not, were quick to absorb main concepts. We could delve deeper and discuss topics such as untrained algorithmic pipelines, local image feature extraction or setup of realistic algorithms including ROC optimization in training.

Most brought their own data sets and used them in the second part of the course.

Everyone was strongly focused on the Deep Learning. People trained their own convolutional nets on imaging use-cases, we have prepared and some also on their own data sets. We have then discussed ways how to use Deep Learning on non-image domains such as ECG processing and text mining.

For us, it is always great to see when some participants push further what they could achieve on their own problems. This time, we even had two researchers, staying a bit longer, managed to come up with the best solutions till date on their respective medical data sets.

All in all, it was an intensive week and a great experience. We also got very valuable feedback that will help us to improve organization of use-case examples. Thank you all!