Deep learning in document analysis
Prof. Thilo Stadelmann, ZHAW School of Engineering
Artificial neural networks have achieved incredible feats in the field of deep learning over the past six years. Not only have they revived public interest in artificial intelligence, their enormous capacities have brought about a paradigm change in pattern recognition and, specifically, image recognition. Today, artificial neural networks outperform humans in many specialised applications.
Professor Thilo Stadelmann’s lecture provides a brief and accessible introduction to deep learning and presents his latest research results in the field of document analysis. Applications such as optical music recognition and newspaper layout analyses may well become commonplace in libraries and archives and simplify the indexing of multimedia holdings significantly.
Thilo Stadelmann is a professor at the ZHAW School of Engineering in Winterthur. His research focuses on artificial intelligence and machine learning. After studying Informatics in Giessen, Germany, he completed his doctorate on multimedia-driven film analysis at the Philipps-Universität Marburg in 2010 as part of a collaborative research centre in the social sciences, whose beneficiaries included archives and libraries. Afterwards, Thilo Stadelmann spent several years in various managerial positions in the automotive industry. His current research explores practical applications of machine learning and, in particular, deep learning for pattern recognition tasks such as document analysis. He is the Scientific Director of ZHAW digital and a co-founder and board member of the Swiss Alliance for Data-Intensive Services.