Deep Learning has proven to be the primary technique to address a number of problems. But each application of AI inevitably encounters unexpected scenarios (edge cases) in which the system does not perform as required. Knowledge-infused learning uses commonsense knowledge encoded in Knowledge Graphs in order to provide capabilities like generalisation, explainability and adaptability of AI systems and thus paves the way towards Artificial General Intelligence.
Can commonsense Knowledge Graphs teach Neural Networks to generalise and explain? - Interview with developer of Knowledge-infused Learning Manas Gaur