SPP 1527:  Autonomous Learning

• Subject Area: Computer Science, Systems and Electrical Engineering ; Biology ; Mathematics ; Medicine ; Physics ; Social and Behavioural Sciences
• Term: from 2011 to 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 172415596

In recent years, computational learning research was tremendously successful in solving many data analysis problems. The methods of machine learning and statistical learning theory have become essential tools in various engineering, life science and natural science disciplines. However, such methods depend to a large degree on an expert to collect the data and represent it in some appropriate format, to decide on a suitable algorithm and hyper-parameters, and to decide on the structure of internal representation. This contradicts the intention that learning should lead to more autonomy and it contrast to learning as we observe it in biological systems.
The aim of this Priority Programme is to develop novel foundations of autonomously learning systems. This calls for new concepts and methods, which go beyond existing machine learning methods, towards systems that autonomously explore an unknown environment and develop appropriate representations. Core aspects of autonomous learning are: (1) the autonomous choice of (hyper-)parameters, representations and features for learning, (2) the autonomous collection of data, i.e., exploration and active search to accelerate learning instead of learning from static data sets, (3) the autonomous development of appropriate representations, including hierarchies, and (4) the incremental abstraction of stimuli, internal representations and actions.
Existing machine learning and robotics methods, in particular reinforcement learning, provide a starting ground for this research. Based on this, we aim for the next step towards truly autonomously learning systems.

• DFG Programme: Priority Programmes
• International Connection: Canada

Projects

• Active exploration in the high-dimensional data of an artificial skin (Applicants Knoll, Alois ;  Strohmayr, Michael )
• An information theoretic approach to autonomous learning of embodied agents (Applicant Ay, Nihat )
• Auto-Tune: Structural Optimization of Machine Learning Frameworks for Large Datasets (Applicants Brox, Thomas ;  Hennig, Philipp ;  Hutter, Ph.D., Frank )
• Autonomous Active Object Learning Through Robot Manipulation (Applicants Behnke, Sven ;  Burgard, Wolfram )
• Autonomous and Efficiently Scalable Deep Learning (Applicant Lücke, Jörg )
• Autonomous Learning for Bayesian Cognitive Robotics (Applicant Beetz, Ph.D., Michael )
• Autonomous Learning of Bipedal Walking Stabilization (Applicant Behnke, Sven )
• Bayesian Learning of a Hierarchical Representation of Language from Raw Speech (Applicant Häb-Umbach, Reinhold )
• Coordination Project SPP 1527 "Autonomous Learning" (Applicant Toussaint, Marc )
• Development as autonomous learning: Emergence of developmental stages that lead from sensori-motor behaviors to embodied cognition (Applicant Schöner, Gregor )
• Efficient Active Online Learning for 3D Reconstruction and Scene Understanding (Applicants Cremers, Daniel ;  Triebel, Rudolph )
• Hyperparameter Learning Across Problems (Applicant Schmidt-Thieme, Lars )
• Learning Dynamic Feedback in Intelligent Tutoring Systems (Applicants Hammer, Barbara ;  Pinkwart, Niels )
• Learning Efficient Sensing for Active Vision (Esensing) (Applicant Martinetz, Thomas )
• Learning Modular Policies for Robot Motor Skills (Applicants Neumann, Gerhard ;  Peters, Ph.D., Jan Reinhard )
• Learning to Behave in the Sensorimotor Loop - An Animat Approach (Applicant Pasemann, Frank )
• Linking metric and symbolic levels in autonomous reinforcement learning (Applicant Obermayer, Klaus )
• Reinforcement Learning with Qualitative Feedback (Applicants Fürnkranz, Johannes ;  Hüllermeier, Eyke )
• Relational exploration, learning and inference - Foundations of autonomous learning in natural environments (Applicants Kersting, Kristian ;  Toussaint, Marc )
• Robots Exploring Tools as Extensions to their Body Autonomously (Applicants Asfour, Tamim ;  Ritter, Helge )
• Scalable Autonomous Reinforcement Learning - From scratch to less and less structure (Applicants Bödecker, Joschka ;  Peters, Ph.D., Jan Reinhard )
• Sparse Coding Approaches to Language Acquisition (Applicant Häb-Umbach, Reinhold )
• The Physical Exploration Challenge: Robots Learning to Discover, Actuate, and Explore Degrees of Freedom in the World (Applicants Brock, Oliver ;  Toussaint, Marc )
• Theoretical concepts for co-adaptive human machine interaction with application to BCI (Applicant Müller, Klaus-Robert )
• Unsupervised Learning of Hierarchical Representations for Natural Images (Applicant Bethge, Matthias )

• Spokesperson: Professor Dr. Marc Toussaint