Code for the paper:
Sun, H., Manchester, W., Jin, M., Liu, Y., & Chen, Y. (2023). Tensor Gaussian Process with Contraction for Multi-Channel Imaging Analysis. ICML 2023.
Files in this repo:
- model.py: code for model class "Tensor-GPST" & "Tensor-GP"
- Simulation_Mod.py: code for generating simulation data
- utils.py: code for other miscellaneous functions (model evaluation, matrix gradient calculation, etc.)
- simulation_code.ipynb: Jupyter Notebook for running a sample model, and running a simulation experiment for model comparison.
-
Run the self-contained notebook simulation_code.ipynb for running our model and simulation.
-
Notes in the Solar Flare Data: Our solar flare imaging dataset is too large to be shared, and there might be conflicts of interest with our collaborators if we release the dataset. We are planning on releasing the dataset together with a scientific journal paper later on. For raw AIA and HMI imaging dataset, one can find the open source version via JSOC.