## Math Calendar

Monday, August 19, 2019

10:30-17:00

km. 611

Tuesday, August 20, 2019

Wednesday, August 21, 2019

Thursday, August 22, 2019

Friday, August 23, 2019

Monday, August 26, 2019

Tuesday, August 27, 2019

09:00-17:00

km.611

Wednesday, August 28, 2019

09:30-17:00

km.611

Thursday, August 29, 2019

Friday, August 30, 2019

Tuesday, September 3, 2019

Tuesday, September 10, 2019

16:00-17:00

HFG 611

Tuesday, September 17, 2019

16:00-17:00

HFG 611

Thursday, September 19, 2019

16:00-17:00

Tuesday, September 24, 2019

Thursday, September 26, 2019

16:00-17:00

Complex Systems Seminar

Rok Cestnik (Potsdam University) - Inferring the dynamics of oscillatory systems using recurrent neural networks, HFG 611

see description

Title: Inferring the dynamics of oscillatory systems using recurrent neural networks

Abstract: We investigate the predictive power of recurrent neural networks for oscillatory systems not only on the attractor but in its vicinity as well. For this, we consider systems perturbed by an external force. This allows us to not merely predict the time evolution of the system but also study its dynamical properties, such as bifurcations, dynamical response curves, characteristic exponents, etc. It is shown that they can be effectively estimated even in some regions of the state space where no input data were given. We consider several different oscillatory examples, including self-sustained, excitatory, time-delay, and chaotic systems. Furthermore, with a statistical analysis, we assess the amount of training data required for effective inference for two common recurrent neural network cells, the long short-term memory and the gated recurrent unit.

Abstract: We investigate the predictive power of recurrent neural networks for oscillatory systems not only on the attractor but in its vicinity as well. For this, we consider systems perturbed by an external force. This allows us to not merely predict the time evolution of the system but also study its dynamical properties, such as bifurcations, dynamical response curves, characteristic exponents, etc. It is shown that they can be effectively estimated even in some regions of the state space where no input data were given. We consider several different oscillatory examples, including self-sustained, excitatory, time-delay, and chaotic systems. Furthermore, with a statistical analysis, we assess the amount of training data required for effective inference for two common recurrent neural network cells, the long short-term memory and the gated recurrent unit.

Friday, September 27, 2019

16:00-17:00

HFG 611

Tuesday, October 1, 2019

16:00-17:00

HFG 611

Thursday, October 3, 2019

Tuesday, October 8, 2019

16:00-17:00

HFG 611

Tuesday, October 15, 2019

Tuesday, October 22, 2019

16:00-17:00

HFG 611

Thursday, October 24, 2019

16:00-17:00

BBG.7.12

Tuesday, October 29, 2019

16:00-17:00

HFG 611

Thursday, October 31, 2019

Tuesday, November 5, 2019

16:00-17:00

HFG 611

Tuesday, November 12, 2019

Thursday, November 14, 2019

Tuesday, November 19, 2019

Thursday, November 21, 2019

16:00-17:00

Utrecht

Tuesday, November 26, 2019

Tuesday, December 3, 2019

Tuesday, December 10, 2019

16:00-17:00

MIN 014

Tuesday, December 17, 2019

Tuesday, January 7, 2020

Tuesday, January 14, 2020

Tuesday, January 21, 2020

Tuesday, January 28, 2020

Thursday, January 30, 2020

15:30-16:30

MIN 201

Tuesday, February 4, 2020

Tuesday, February 11, 2020