Caldera 🌋

PLEASE NOTE PROJECT HAS MOVED TO Caldera2

Caldera is a Pytorch extension for learning on graphical data. It learns underlying causal relationships in data.

Gallery

Usage

Installation

via pip

via Conda

via Docker

Build CUDA enable image

docker build . -f docker/cu101/Dockerfile

Build slim cpu-only image

docker build . -f docker/cpu/Dockerfile

Tour

Acknowledgements

• rusty1s/pytorch_scatter - an awesome library that provides missing scatter methods in pytorch. Necessary for graph attribute aggregation
• PyTorch Geometric -
• PGN - provided a blueprint for implementing this library

References

Inspirations

• Pytorch Geometric [ADD REF]
• GraphNets [ADD REF]
• Pyro Probabilistic Programming Language
• Pearl, J. (2009). Causality: Models, Reasoning and Inference.
• Pearl, J. (2018). Theoretical impediments to machine learning with seven sparks from the causal revolution.