Abstract: AI is about building machines that act intelligently. The general problem solving tools that have resulted from this pursuit have given us the ability to solve previously unsolvable problems.
Simultaneously, important problems where the only current viable solution is human intuition, formed from years of experience, are only increasing in frequency and complexity. By formulating these problems to be solved by AI techniques, we gain dramatic increases in scale and efficiency, and sometimes solve problems that were previously not solvable.
In practice, we have found often there is a large part of the problem that is not analytically expressible, and only exists within the mind of the human operator. By positioning the human operator to work along side an interactive AI, this knowledge can be captured, using techniques such as deep learning. The result is practical solutions to problems, the complexity of which have previously overpowered both humans and machines.
In this talk, I will discuss two specific domain examples I have been involved with, in which an interactive AI was positioned to work collaboratively with domain experts. In each case, I will talk about how as a result, the human was able to perform their job at a dramatically higher level of efficiency, and in some cases solve problems previously not possible.
Bio: Don M. Dini has been practicing, teaching, and writing about data science and intelligent systems for over ten years. He studied computer science and artificial intelligence at University of Illinois at Urbana-Champaign and University of Southern California. While at USC he was a lecturer in computer science and worked on applying AI to various real world problems, such as understanding city populations through simulation, and systems to provide security against unknown attackers, which have since been used at LAX, the US coast guard, among other institutions. At AT&T he was a principal data scientist, and applied AI to manage the complexity of diagnosing and operating modern communication networks. Today Don is director of AI and Automation at foodRev, a non-profit that uses technology to distribute food to the hungry. Don is also a data science lecturer at Udacity, where he co-created the course, Model Building and Validation.