Fairer Software Made Easier (using "Keys")

TL;DR

This paper explores leveraging the 'keys effect' in software systems, where a few key features dominate, to simplify explanations and control, potentially improving transparency and ethical AI management.

Contribution

It introduces the concept of the 'keys effect' and demonstrates how exploiting it can make complex software explanations more manageable.

Findings

Systems often exhibit a 'keys effect' with few features controlling the rest

Exploiting keys allows for simplified 'what-if' analysis

Potential to improve ethical AI explanations

Abstract

Can we simplify explanations for software analytics? Maybe. Recent results show that systems often exhibit a "keys effect"; i.e. a few key features control the rest. Just to say the obvious, for systems controlled by a few keys, explanation and control is just a matter of running a handful of "what-if" queries across the keys. By exploiting the keys effect, it should be possible to dramatically simplify even complex explanations, such as those required for ethical AI systems.