adaamko / POTATO

XAI based human-in-the-loop framework for automatic rule-learning.
MIT License
47 stars 8 forks source link

Training command rules #64

Closed GKingA closed 2 years ago

adaamko commented 2 years ago

Thanks for the PR @GKingA. Could you please write a few sentences about the changes in the graph_extractor as currently I'm not sure I understand fully what is changed and why. Thank you!

GKingA commented 2 years ago

Hi @adaamko! The previous logic was based on subgraph mapping and gathering the possible good fits for the .* in a rule. This version uses the same logic, but does so with a GraphMatcher, so the new matcher commands can also be used as the input of the training.