Lamarr Project

The Lamarr Institute for Machine Learning and Artificial Intelligence 

LAMARR Institute of Machine Learning and Artificial Intelligence strives to create the future of European AI. Constituted by the Technical University Dortmund, University Bonn, Fraunhofer IAIS, and Fraunhofer IML, researchers at LAMARR focus on the research and development of high-performing, trustworthy, and sustainable AI solutions for hands-on business applications. The critical research interest areas are Hybrid Machine Learning, Resource-aware Machine Learning, Trustworthy Artificial Intelligence, Embodied Artificial Intelligence, and Human-centred Systems, focusing on applications such as Planning and Logistics, Astrophysics, Industry and Production, Life Sciences, and Natural Language Processing. 

FLW Lamarr Projects: 

Shareable Human Activity Data for Logistic  

The future of Logistics is bound to be technology and artificial-intelligence-intensive. The idea of seamless collaboration between Man and Robotics can be visioned as the near future of Logistics. However, most Machine Learning methods are data-intensive, and generating Human Activity data, especially in Logistics, may affect human employees’ performance and privacy concerns. The idea of shareable human activity data for Logistics aims to generate unbiased human activity data for varied logistics scenarios with the support of AI.  

DoUnseen library: Zero-shot object detection for robotic grasping: 

The goal of our DoUnseen library is the fast deployment of deep learning in robotic grasping. The library is built on top of two main modules, “unseen object segmentation” and “zero-shot classification”. Combined, the two modules enables the direct deployment of deep learning in object detection application with no need for data collection, retraining, or fine-tuning. Objects can simply be detected by capturing a few images of them and adding them to a database. 

Library: https://github.com/AnasIbrahim/image_agnostic_segmentation 
Homepage: https://lamarr-institute.org/

Gefördert durch: Bundesministerium für Bildung und Forschung 
Ansprechpartner: Anas Gouda, M.Sc., Nilah Nair, M.Sc.
Lamarr-Institute