Knowledge graphs are enjoying a renaissance in AI research. Perhaps due to their success in Game AI and Goal Oriented Action Planning. And this academic paper g

Knowledge graphs are enjoying a renaissance in AI research. Perhaps due to their success in Game AI and Goal Oriented Action Planning. And this academic paper graph looked like one of the more "pure" datasets. The quality of the agent depends on the input clearly. And machine understanding of academic research is certainly a problem rich in potential rewards. Linked below are the results of a challenge to "solve" a Sherlock Holmes mystery based only on the text! One of the entries involved constructing a knowledge graph of all potential suspects in the entire Conan Doyle corpus ;) Report on the First Knowledge Graph Reasoning Challenge 2018 -- Toward the eXplainable AI System https://arxiv.org/abs/1908.08184 https://arxiv.org/abs/1908.08184 Microsoft Research: Explainable AI for Science and Medicine https://www.youtube.com/watch?v=B-c8tIgchu0 https://www.youtube.com/watch?v=B-c8tIgchu0 SHAP (SHapley Additive exPlanations) is a unified approach to explain the output of any machine learning model https://github.com/slundberg/shap https://github.com/slundberg/shap