chizson
UNDER CONSTRUCTION
CHIZSON is a Python-based Neuro-symbolic framework for knowledge and neural networks integration. The framework aims to support AI users and developers to easily represent and incorporate symbolic knowledge and logic programs to connectionist machine learning models. I work on this framework during my free time so the maintainance will be slow. If there are any questions, please drop me and email at: sontn.fz@gmail.com.
User Guide
- Step 01: Prepare a knowlege base
- Step 02: Choose a predicate
- Step 03: Choose a learning method (or load pre-trained models)
- Step 04: Choose a reasoning method
- Step 05: Run
Examples
List of examples
- CompareKB
- AdditionKB
- Part-of
Tutorial
Neural Predicates
Symbolic Neural Predicate
- Auto-Encoder NP
- RBM Predicate
Predictive Neural Predicate
Generative Neural Predicate
General Neural Predicate
- Compositional Neural Predicate
- DBN Predicate
Reasoning
Chaining
- Voted Backward-Forward Chaining
Best Satisfiability
- Gibbs Sampling Reasoner
- Free-Energy Minimiser