2 years PostDoc Zero-Example Deep Learning for Vision at University of Amsterdam

A postdoctoral research fellow position in Zero-Example Deep Learning is open in the Informatics Institute of the University of Amsterdam (the Netherlands) under guidance of Dr. Thomas Mensink, starting as soon as possible.
Project description
The goal of the research position is to study zero-example learning, that is learning without (explicit) visual examples, by using inter-mediate visual representations and reasoning over the visual extend of novel categories. Most current zero-shot prediction systems, use deep convolutional network as feature extractor, yet do not exploit the knowledge encoded in the deep network further. The aim of this project is to devise methods for zero-shot learning directly using transfer within the deep network. The aim is to work on both image classification and video recognition.
Requirements
PhD in Computer Science, emphasizing computer vision and machine learning;
research record in computer vision, video categorization, zero-shot prediction, semantic embeddings, and/or deep learning is considered a pre;
strong publication record in top-tier international conferences and journals;
solid knowledge of programming (for large-scale processing);
assist in guiding PhD and MSc students working on computer vision topics.
See the full application procedure here: http://www.uva.nl/en/about-the-uva/working-AT-the-...