Track: Emerging Analytics
Data is Dirty: What AI Algorithms Actually Do and Why We Need Explainable Artificial Intelligence
Wednesday, April 14, 10-10:40am EDT
Modern black-box artificial intelligence algorithms are computationally powerful yet fallible in unpredictable ways. We argue that Explainable Artificial Intelligence (XAI) frameworksare required to not only understand ‘why’ AIs make their decisions, but to also integrate with human operators so these explanations provide the right quantity and quality of information to each operator. In essence, we need to account the expertise and goals of the user in order for these algorithms gain widespread adoptance in mission-critical environments. In this talk, we will present a survey of introspection techniques as well as a set of requirements for robust XAI systems. We describe a use-case for cognitive models to act as a bridge between humans and AIs that meet most of these requirements.