Selected publications

Home | CV | Publications |

For a complete list of publications click here

On recurrent neural networks

In recent work, we investigated recurrent neural networks for stochastic dynamic system identification and noisy time series prediction. We consider recurrent networks as nonlinear state space models and present a training algorithm based on Expectation-Maximization. A paper was accepted for oral presentation at ICANN 2009:

J. Unkelbach, Sun Yi and J. Schmidhuber; An EM based training algorithm for recurrent neural networks, ICANN , 2009 (pdf)

On optimization in radiation therapy planning

Extensive research has been carried out in order to incorporate various types of uncertainty into treatment plan optimization for radiation therapy with protons and high energy X-rays. Stochastic and robust programming methods have been customized for this task, leading to numerous journal publications and conference proceedings:

J. Unkelbach, B. C. Martin, M. Soukup and T. Bortfeld; Reducing the sensitivity of IMPT treatment plans to setup errors and range uncertainties via probabilistic treatment planning, Medical Physics, 36, p149-163, 2009

J. Unkelbach, T. C. Y. Chan and T. Bortfeld; Accounting for range uncertainties in the optimization of intensity modulated proton therapy, Physics in Med. Biol., 52, p2755-2773, 2007

J. Unkelbach and U. Oelfke; Inclusion of organ movements in IMRT treatment planning via inverse planning based on probability distributions, Physics in Med. Biol., 49, p4005-4029, 2004

PhD thesis

The thesis describes methods to account for organ movements in the optimization of radiotherapy treatment plans, with applications to respiratory motion in the lung and day-to-day movements of the prostate:

J. Unkelbach; Inclusion of organ motion into IMRT optimization using probabilistic treatment planning, PhD thesis, University of Heidelberg, Germany, 2006 (pdf)