Explainable AI - The Apex of Human and Machine Learning

Explore the concept of Explainable AI in this 42-minute conference talk from Strange Loop. Delve into the limitations of Black Box AI technologies and the emerging trend of Explainable AI designed to be intuitive and understandable to human users. Learn about the fundamental differences between human and machine learning, and discover research at the intersection of computation and cognition that has led to machines capable of efficiently communicating their knowledge. Gain insights into building Explainable AI, its applications in safety-intensive and scientific domains, and how it addresses the unexpected failures of Black Box AI. Examine topics such as image classification, problems with explanation, teaching image categories, recursive math, and DARPA's involvement in this field.

Introduction
Problems with mainstream AI
AI for highrisk tasks
Explainable AI
Image Classification
Problems with Explanation
How do we fix these problems
Teaching image categories
Recursive math
DARPA