mossjacob / alfi-paper

Alfi: Approximate Latent Force Inference

ARCHIVE: Current, maintained release in mossjacob/alfi

Please refer to https://github.com/mossjacob/alfi for the up-to-date code and examples.

---

---

Implement Latent Force Models in under 10 lines of code!

This library implements several Latent Force Models. These are all implemented in building blocks simplifying the creation of novel models or combinations.

We support analytical (exact) inference in addition to inducing point approximations for non-linear LFMs with ODEs and PDEs written in PyTorch.

Documentation

See Jupyter notebooks in the docs/notebooks/ directory. The docs contain linear, non-linear (both variational and MCMC methods), and partial Latent Force Models. The notebooks also contain complete examples from the literature such as a replication of the analytical linear Latent Force Model from Lawrence et al., 2006

These notebooks should help get started with LaFoMo.

Linear LFMs

For analytical Latent Force Models, have a look at these notebooks carrying out exact inference in the directory:

notebooks/linear/*

Non-linear LFMs

For non-linear Latent Force Models, the variational approach is used, in the directory:

notebooks/nonlinear/*

Partial Differential Equations

notebooks/pde/*

Neural Latent Force Operator

notebooks/nn/*

alfi-paper's Issues

Missing files

Hi, first of all thank you so much for providing the code of your paper.

I am interested in replicating the results from Lawrence et al., 2006 and I was looking at the implementation in https://github.com/anony-submissions/alfi/blob/main/notebooks/linear/exact.ipynb

It appears that some files are missing, in the notebook these imports are present:

from alfi.models import ExactLFM
from alfi.trainers import ExactTrainer
from alfi.datasets import P53Data
from alfi.plot import Plotter1d
from alfi.utilities.data import p53_ground_truth

but in the alfi folder I can only find datasets, kernels, and mcmc. This means that models, trainers, plot and utilities are missing.

Could you help my find the rest of the code?

Thanks in advance,
Andrea