Humans, Data and Machines

Jane Bambauer, Professor of Law, University of Arizona James E. Rogers College of Law

Machine learning is shaping human lives in both obvious and subtle ways. Important economic and legal decisions about credit, employment, and criminal justice are already made with the aid of complex algorithms, raising difficult questions about whether machines can make decisions that are accurate and fair. Machine learners can become biased when the programmed objectives or the training data used to teach the algorithm are flawed. On the other hand, machines have some advantages over humans since they do not apply pre-existing assumptions and can more quickly recognize unexpected patterns. Machine learning also affects the human experience by creating advertising, suggestions, chat-bots, and even auto-generated news articles tailored to the individual. The government has some power to constrain artificial intelligence, but there are practical and constitutional limits to legal interventions.

Vincent J. Del Casino Jr., Vice President, Academic Initiatives and Student Success, Professor, School of Geography and Development, University of Arizona

This talk challenges the notion that “big data” are what people believe they are – large, singular inanimate manifestations of our proxy selves – and argues that there is no “big data” really, just millions of small bits and pieces brought together through a series of algorithmic possibilities. But, big data analytics and the robotic futures that they engender are clearly producing anxieties for everyday social life and institutions, such as the university, have to manage these anxieties as they rethink themselves in relation to big data analytics and their concomitant robotic futures. As a result, universities have to double-down on investments in a broad education by asking how big data are represented in society, how human life is being organized in relation to big data, and how an interdisciplinary future can help manage the rapid changes produced by advances in robots and robotic technologies.

Luis von Ahn, CEO and Co-Founder, Duolingo, Professor of Computer Science, Carnegie Mellon University

This talk is about harnessing human time and energy to address problems that computers cannot yet solve. Although computers have advanced dramatically in many respects over the last 50 years, they still do not possess the basic conceptual intelligence or perceptual capabilities that most humans take for granted. By leveraging human skills and abilities in a novel way, I want to solve large-scale computational problems and collect training data to teach computers many of the basic human talents. To this end, I treat human brains as processors in a distributed system, each performing a small part of a massive computation

Nirav Merchant, Director Data Science Institute, Data7, University of Arizona

Machine learning (ML) based systems are rapidly becoming pervasive, powering many applications from recommending music, movies and merchandise to driving our cars to assisting in medical diagnoses. Our daily interactions, behavior, and choices, whether we are aware of them or not, are the sources of data for training these systems. But how are these ML based platforms built and utilized ?. While ML based platforms create amazing opportunities, especially when coupled with advances in cloud computing, reliance on these platforms comes with ethical, security, and technical concerns. How do we strike a balance for enabling pragmatic and productive use of these capabilities? ML powered platforms are gaining proficiency and becoming deeply integrated into existing and emerging automation across many domains of science and society, causing a shift in opportunities impacting many professions. What are the new learning and training opportunities that allow us to stay relevant and lead the way for future innovations

Mihai Surdeanu, Associate Professor of Computer Science, University of Arizona

We are inundated daily with news about artificial intelligence (AI) achieving tremendous results, e.g., defeating human champions at Go, driving better than us, etc. But does this mean that we are approaching the technical singularity where artificial intelligence far surpasses the human one? Does this mean that machines truly think? In this talk we will analyze these questions and illustrate that AI does not think that way we think: machines do not have a good way to represent and reason with world knowledge, and, of course, they are not self aware. Instead, AI is designed to automate and scale up pattern recognition for specific tasks. Because of this different goal, AI does perform better than humans at certain tasks. I will review a series of problems where AI outperforms humans, including specific applications of natural language understanding, precision medicine, identifying planetary objects, and other problems, many of which implemented here at University of Arizona.

Stephen Kobourov, Professor of Computer Science, University of Arizona

The idea of computation and algorithms is old, but modern day computers are a relatively new phenomenon. Even more recent are the notions of artificial intelligence (AI), machine learning (ML) and big data. While it is difficult to clearly define AI and ML it is evident that progress in these fields, combined with access to large datasets, has a significant impact on all aspects of our lives. This raises new mathematical and engineering challenges (can we solve previously unsolvable problems?), but also philosophical questions (can machines think?), and considerations in ethics and law (can machines be more objective than humans?).