// Dr. rer. nat. Frank Sehnke
Research Departments
Systems Analysis
Contact
+49 711 78 70-303
Career
Since 2011 Researcher in Machine Learning with focus on Deep Neural Networks (DNN) and Optimisation at the ZSW
2008 to 2012: Part of the machine learning library PyBrain development team. (www.pybrain.org)
2007 to 2010: Ph.D. student of Cognitive Robotics Group at the TU Munich Cogbotlab of the TUM Computer Science department.
Study of Computer Science at the University of Tübingen, Germany, till 2005
2001 to2006: Part of the RoboCup team Attempto Tübingen
Fields of activity
- Wind-, Photovoltaic-, Waterpower-forecasting using DNNs with state of the art learning methods like RBM pretraining, RMSProp-, Adam-Training and Automatic Feature Selection
- Atmospheric Ozone predictions with DNNs from satellite UV and IR spectral data
- Windpotential Site Assessment using DNNs and statistical methods for Windspeed Distribution conservation
- Main developer of the Framework P2IONEER: Optimisation of Energy-Systems with conventional and renewable energy producers with Parameter-based Policy Gradients
- Optimisation of production systems with Model-Based Reinforcement Learning and Evolutionary Algorithms using DNNs and Gaussian Processes
- Computer Vision for Cloud Motion prediction for Photovoltaic Now-Casts.
Selection of Publications
- Frank Sehnke, Christian Osendorfer, Thomas Rückstieß, Alex Graves, Jan Peters, and Jürgen Schmidhuber Parameter-exploring policy gradients. Neural Networks, 23(2), 2010.
- Frank Sehnke, Tingting Zhao Baseline-Free Sampling in Parameter Exploring Policy Gradients: Super Symmetric PGPE Artificial Neural Networks, Volume 4 2015, Pages 271-293
- Frank Sehnke, Achim Strunk, Martin Felder, Joris Brombach, Anton Kaifel, Jon Meis Wind Power Resource Estimation with Deep Neural Networks.Artificial Neural Networks and Machine Learning–ICANN 2013
- Frank Sehnke Parameter Exploring Policy Gradients and their Implications.PhD-Thesis, Technical University Munich 2012
- Frank Sehnke, Martin Felder, Anton Kaifel Learn-O-Matic: A Fully Automated Machine Learning Suite for Profile Retrieval Applications.EUMETSAT Meteorological Satellite Conference 2012
- For further publications, please refer to my Google Scholar page.