Deep-PTL: Life-long learning with flexible data modalities for automation
in production, transport, and logistics

September 2018 - August 2021

Project members

Prof. Dr. Thomas Brox
Sudhanshu Mittal

Abstract

The project develops new deep learning methods to automate processes that have been controlled manually so far, and to make automated processes more reliable, more efficient, and more flexible. Special focus is given to the following challenges in machine learning: life-long learning, high-variability of input data, estimation of the system's uncertainty, and scalability to embedded platforms. The benefit of the developed methods will be demonstrated in dedicated application domains in production, transport, and logistics.

Publications