Python + PyTorch Reproduction of Experiment 4.5 of the Discriminative Kalman Filter paper from 2020.
NOTE: this code has not been run on a Mac architecture. I ran everything on a PC, so I recommend using x86_64 architectures.
See requirements.txt for package dependencies.
First, run flint_preprocess_data.ipynb using the instructions in its README.
Next, run Paper_Script_45.ipynb to produce the first two results tables.
To produce the last two results tables, run Paper_Script_45.ipynb with nine additional different seed values, then run Paper_additional_tables.ipynb.