Jan Melchior

Background

About Jan Melchior

My PhD in machine learning and computational neuroscience has been about artificial neural networks and deep learning, in particular on improving the training of unsupervised neural networks as well as a better mathematical understanding of various training algorithms such as back-propagation, contrastive learning, as well as Hebbian learning. As a computer scientist my passion is coding, which I intensively followed during my research that led to the python machine-learning library named PyDeep, which allows others to reproduce my research results easily. I enjoyed teaching and supervising several bachelor and master thesis and always prefered working in a team. Besides I worked for IT-companies during my bachelor, master, and PhD studies. This gave me the opportunity to implement and use various machine learning algorithms to successfully solve problems such as vehicle replacement forecasting, classification of invoice positions, image classification, object detection, semantic segmentation, and fuzzy string search.