Configuration and Visualization of Deep Neural Networks for Document Retrieval
Deep neural networks are currently the leading technique for machine learning tasks like document retrieval, which is the basis of every search engine. The main benefit of deep neural networks in comparison to other techniques is that the time needed for retrieving a document is independent of the amount of available documents. However, it is often difficult to configure a deep neural network. The goal of this project is to build a tool which eases the configuration of traini
Team: Leon Schröder
Studiengang: Computer Science and Media (Master)
Betreuer: Prof. Dr. Johannes Maucher