Quantum Information & Computing Meeting: Forecasting Chaotic Dynamics with Machine Learning
 

Cosimo Rusconi (MPQ):
I will present the paper PRL 120, 024102 (2018) (see also article cover at QuantaMagazine) where the authors implement a recurrent neural network (RNN) which is able to predict the dynamics of a classical chaotic system several Lyapunov time in the future after training on one input solution.

Cosimo Rusconi (MPQ)
Group meeting (hybrid format: online/seminar room B2.46)
Tue, 7 December 2021, 11:30 am (MEZ)

Abstract:

I will present the paper PRL 120, 024102 (2018) (see also article cover at QuantaMagazine) where the authors implement a recurrent neural network (RNN) which is able to predict the dynamics of a classical chaotic system several Lyapunov time in the future after training on one input solution. I will introduce the Machine Learning Model they use and explain the main motivations. If there will be time, I will compare the approach followed by the authors with other approaches in the literature where people apply machine learning to (quantum) physics problem underlining differences and similarities.