This project aims to build a general framework for DataLog neural program systhesis.
Now, two backends are implemented:
- ∂ILP in "Learning Explanatory Rules from Noisy Data": https://www.jair.org/index.php/jair/article/view/11172
- NTP in "End-to-End Differentiable Proving"
- note that the official implementation of NTP can be found in https://github.com/uclmr/ntp
- numpy
- tensorflow (>=1.7)
- run main.py for predecessor problem