DFG Grants University of Bremen a Further Collaborative Research Center

“Lay the table!” – executing simple instructions like this might not be so difficult for humans to perform: but for robots they present an enormous challenge. Carrying out even simple tasks takes whole lot of extremely complex information for granted: What precisely should be laid on the table?? Which is the best way to hold a glass? And even how to know what a table is in the first place! In order that robots are able to autonomously follow instructions involving a certain degree of abstraction, they must first be enabled to obtain the missing information themselves. The German Research Foundation (DFG) has just granted the University of Bremen a further collaborative Research Center (CRC) in which scientists will apply completely new approaches to “training” robots to independently execute such abstract instructions: In other words, they are helping them to “learn”. July 1, 2017, will see the start of the CRC EASE (Everyday Activity Science and Engineering). The initial four-year funding phase it is being financed by DFG in an amount of 10 million euro.

Strengthening the high-profile research area “Minds, Media, Machines”

The President of the University of Bremen, Professor Bernd Scholz-Reiter, commented the news by saying: “The DFG’s support for the new CRC provides further confirmation that our high-profile research area ‘Minds, Media, Machines’ creates strong synergizing effects by combining the expertise of the research fields of artificial intelligence, cognitive systems, and robotics. I’m delighted at the news and my congratulations go to the whole team”. The leader of the University’s high-profile research area, Professor Kerstin Schill, who is also the CRC’s deputy spokesperson, pointed out that “The CRC will further strengthen basic research, teaching and internationalization in the whole high-profile area”.

Robots should increase the quality of life

In the framework of basic research, EASE will investigate how humans are able to arrange everyday activities with flexibility, efficiency, and reliability. Such an understanding is intended to inspire a new generation of models for robotic control.  “Developing robots to the point that they are able to follow instructions put in simple language opens up a whole range of possibilities to improve life quality – for people with disabilities, for example, or senior citizens who want to live independently in their own four walls for as long as possible”, explains the new CRC’s spokesperson, Professor Michael Beetz. Researchers can think of many more versatile applications for intelligent robots such as carrying out experiments with hazardous materials, for example.

Democratization of robotics

The CRC is of high societal relevance in other aspects, too: It brings together international scholars and scientists and supports the idea of “Open Research” – the concept of making research results easily available to anyone who has an interest in knowing them. It therefore systematically pursues the aim of democratization in the field of robotics and artificial intelligence so that developments will in future not be dominated solely by commercial interests. Working in cooperation with other universities, the University of Bremen has been laying the groundwork for intensive collaboration in the field of robotics for a number of years now.

Interdisciplinary research

Research activities in the CRC are highly interdisciplinary, involving not only computer science and mathematics, but also linguistics as well as the human and health sciences. Other external cooperation partners include the Excellence Cluster Cognitive Interaction Technology (CITEC) at the University of Bielefeld (Professor Helge Ritter), the Institute of Robotics and Mechatronics at the German Aerospace Center (Professor Alin Albu-Schäffer), and the Chair for Cognitive Systems of the Technical University of Munich (Professor Gordon Cheng).

Three main research areas

The CRC is divided into three main research areas, each of which deals with a specific aspect of everyday human activity. In the first area, ‘descriptive models of human everyday activity’, the EASE researchers develop new models for describing everyday human activity using the methods of neuropsychology, neuroinformatics and linguistics. The second area, ‘principles of information processing for everyday activity’, among other things investigates new methods of science presentation and automatic deductive reasoning for everyday activities using methods of theoretical informatics. In the third research area, ‘generative models for mastering everyday activity and their embodiment’, the insights gained from the two other areas are transferred to robotic systems and subsequently put to the test.

“Household marathon” for robots 

By the end of the initial funding phase, the robots should be able to autonomously take part in a veritable “household marathon” comprising the performance of tasks like laying the table and preparing simple meals over longish periods of time.

If you would like to have more information on this topic, feel free to contact:

University of Bremen
Prof. Michael Beetz, PhD (CRC Spokesperson)
Phone: +49 (0) 421 218 64000
email: ai-officeprotect me ?!cs.uni-bremenprotect me ?!.de 

Mann steht neben drei Robotern.
Artificial Intelligence Laboratory. Picture by Alexis Maldonado (Oct-2016). Copyright Universitaet Bremen