Notebooks for a project with aim of predicting the response of Earth's external magnetic field to solar activity via ''deep learning''.
Goals/To do:
- First zonal harmonic prediction with basic LSTM network, forecasting_q10_LSTM
- Dst prediction with basic LSTM network (to compare with previous results), forecasting_Dst_LSTM
- Implement Garnelo et al.'s Bayesian neural network architecture and apply to first zonal harmonic prediction
- Train networks only on storm time series
- Consider subsets of input OMNI data to train on and compare effectiveness of corresponding networks
- Interrogate neurons of trained networks to elucidate their sensitivities to particular inputs and thereby try to uncover whether or not a physical model was approximated by the network (e.g. did the network construct any terms analogous to those outlined by Burton et al. 1975?)
Note to self: first ssh @ela.cscs.ch then ssh daint