Good overview of the challenge. We can now solve small parts, but then it is a craft making those solutions work together in context and practice. It is a challenge. Include your decision makers, people with knowledge about the data and business practices. Involve them early and often. Consider context and consequences. Mold process scripts and have conversations about how they work. Keep testing and be ready to adapt.
Introductory video:
The Challenge of Crafting Intelligible Intelligence By Daniel S. Weld, Gagan Bansal
Communications of the ACM, June 2019, Vol. 62 No. 6, Pages 70-79
10.1145/3282486
Artificial Intelligence (ai) systems have reached or exceeded human performance for many circumscribed tasks. As a result, they are increasingly deployed in mission-critical roles, such as credit scoring, predicting if a bail candidate will commit another crime, selecting the news we read on social networks, and self-driving cars. Unlike other mission-critical software, extraordinarily complex AI systems are difficult to test: AI decisions are context specific and often based on thousands or millions of factors. Typically, AI behaviors are generated by searching vast action spaces or learned by the opaque optimization of mammoth neural networks operating over prodigious amounts of training data. Almost by definition, no clear-cut method can accomplish these AI tasks. ... "
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment