IMPORTANT NOTE This project is deprecated as of 11/2024 and is being
merged with resp_protein_toolkit.
All of the functionality available in this repo is now available in
resp_protein_toolkit
which can be easily installed using pip and contains
significant additional functionality not available in this project. We recommend
using resp_protein_toolkit
in future.
A PyTorch class that implements an approximate Gaussian process as the last layer of a neural network - compatible with any architecture and with regression, binary logistic classification and classification. This provides a simple way to obtain uncertainty calibration.
We recommend using this in combination with spectral normalization which is approximately distance-preserving (see Liu et al for details). This ensures that datapoints far from the training set in the input space are appropriately associated with high uncertainty. We may add standard spectral-normalized layers to a future release to make this easier to implement.
For details on installation and usage, see the docs