25 May 2020 - Kobby Nuamah

Link

https://eu.bbcollab.com/guest/97e5b2666aa74b30870a4c83ad03760a  

Speaker

Kobby Nuamah

Title

 Explainable Inference in the FRANK Query Answering System

Abstract

The demand for insights into how artificial intelligent systems work is rapidly growing. This has arisen as AI systems are being integrated into almost every aspect of our lives from finance to health, security and our social lives. Recent techniques for generating explanations focus mostly on explaining opaque algorithms, such as the output of a neural network model. However, such models do not work in isolation; they are combined, either manually or automatically, with other inference operations. Hence, local explanations of individual components are not enough to give the user adequate insights into how an intelligent system works.

Also, it is not unusual for a system made up of fairly intuitive components to become opaque when it is combined with others to build an intelligent agent.

We argue that there is the need to combine diverse forms of reasoning in order to generate explanations that span the entire chain of reasoning.

In this talk, I'll give a brief introduction to the FRANK (Functional Reasoning for Acquiring Novel Knowledge) question-answering system and then discuss our use of "explanation blankets"  to get insights into how FRANK infers its answers. I will also show how interacting with FRANK's inference graphs enhances the insights it provides.