“These days, it’s almost impossible not to hear about AI every day.”
So began day two of the inaugural AI Innovations Forum held Nov. 4-5 at SAS world headquarters in Cary, N.C. Udo Sglavo, vice president of Analytics R&D for SAS, introduced the full day of presentations revealing the newest research on AI and showcasing the latest and greatest applications of machine learning, deep learning and the like by pointing out how ubiquitous and commonplace AI has become in our lives.
While the Nov. 4 event consisted of a series of hands-on workshops for students and educators, the Nov. 5 session gave attendees insights into recent developments in AI and machine learning from world-renowned leaders in the field. The forum, co-hosted by the Kenan Institute of Private Enterprise and analytics leader SAS, featured such experts as Peter Guerra, North America chief data scientist for Accenture; Thomas Zerdick, head of unit IT policy for the Office of the European Data Protection Supervisor; and Jianqing Fan, Frederick L. Moore Professor of Finance and director of financial econometrics and statistics at Princeton University.
Fan’s keynote helped set the stage for the day, with definitions of AI, machine learning and deep learning. Presenting a brief history of AI from when the term was coined in 1955, Fan noted that AI is achieving ever-greater success due to the large-scale optimization big data and modern computing power make possible. The ideal, said Fan, is to create “humanoid” machines with cognition, thinking and perception. Although challenges still exist, he added, “we are at the reality now of [developing these machines], liberating humans from tedious and dangerous jobs.”
Zerdick focused on one of AI’s greatest challenges – data privacy and security. His keynote outlined the European Union’s General Data Protection Regulation (GDPR), considered by many to be the gold standard for data protection measures. Zerdick acknowledged that in Europe, as elsewhere, many people suffer from “AI angst,” or distrust of AI. Noting that the U.S. and China lead global research and production of AI-based technologies but lag behind Europe in developing comprehensive data protection rules, Zerdick advocated for a global, human-centered approach toward AI and data privacy. “Computers and technology should serve humans,” said Zerdick, “and humans are responsible for ensuring that.”
Another forum keynote featured Melissa Strong, founder and lead data scientist for IndiOmics. Strong focused on the use of AI in individualized disease prevention and treatment. She described how IndiOmics used machine learning to establish correlations between factors such as exposure to environmental toxins, waist circumference, marital status, income level and cancer diagnoses. The company is also using AI to generate data to investigate the feasibility of individualized breast cancer therapies. A great advantage of using AI simulations for such experimentation, said Strong, is that AI offers huge time and cost savings over conventional lab-based studies.
The day’s breakout sessions covered many aspects and uses of AI, from evaluating and maximizing the cost-benefit ratio of a major catalog retailer’s direct mail marketing to monitoring and incentivizing employee behaviors at a well-known banking institution to using neural networks to develop investment portfolio strategies. Daniel Ringel, assistant professor of marketing at UNC Kenan-Flagler Business School, presented a case study in which Home Depot analyzed the sentiment in customer social media posts to gauge shoppers’ satisfaction and predict the company’s future performance based on that sentiment. In all, attendees enjoyed nearly 20 presentations designed to help spark innovation in their own organizations.
The SAS AI Innovations Forum was co-chaired by Brett Wujek, principal data scientist with the artificial intelligence and machine learning R&D division at SAS, and Professor Eric Ghysels, faculty director of the Kenan Institute’s Rethinc. Fintech Lab. To learn more about the Kenan Institute’s work with emerging technologies, please click here.