Autonomous Learning

From 2012-2018 the German Research Foundation supports basic research on learning methods in artificial systems with the Priority Programme 1527 Autonomous Learning. More than 50 scientists from the fields of machine learning, robotics and neuroscience participate in the programme (see map and summary of the scientific programme below, a complete list of all projects here).

The special issue Advances in Autonomous Learning of the Künstliche Intelligenz journal was published by Springer-Verlag in November 2015. Most of the articles are contributions from our priority programme participants, which you can read here. To read all articles as published by Springer, follow the link below.


Download the editorial by Barbara Hammer and Marc Toussaint here.

Summary of the scientific programme

Autonomous Learning research aims at understanding how autonomous systems can efficiently learn from the interaction with the environment, especially by having an integrated approach to decision making and learning, allowing systems to autonomously decide on actions, representations, hyperparameters and model structures for the purpose of efficient learning. Specific research topics are:

  • Research that aims to make machine learning methods more autonomous w.r.t. to the choice of hyper-parameters, representations, kernels, and specific learning algorithm. Research on learning deep representations.
  • Systems that have learning as their primary goal, as in active learning, autonomous exploration and intrinsic motivation approaches.
  • Research on the above in the concrete robotics context, that is,integrated active learning and decision making in natural environments.
  • Mathematical views on analysis, modelling and evaluation of autonomy itself and autonomously learning systems.

Initiators of the Priority Programme proposal

See also Initiators.

Coordination & information

Marc Toussaint, Alica Abberger & Sophie Schroth
Stuttgart University, Universitätsstraße 38, 70569 Stuttgart, Germany

Funded projects

The maps shows all participating projects.

Further information: