Sign In

Communications of the ACM

ACM TechNews

Looking to the Future of Data Science

View as: Print Mobile App Share:
Oren Etzioni and Eric Horvitz

Allen Institute's Oren Etzioni (left) and Microsoft Research's Eric Horvitz (right) presented strikingly different views of big data in separate KDD2014 keynotes.

Credit: Geekwire, Microsoft

Dueling keynote speeches opening the first two days of ACM's Knowledge Discovery and Data Mining conference in New York this week demonstrated different visions of what the future of data science should look like.

In his Monday keynote, Oren Etzioni, chief executive of the Allen Institute for Artificial Intelligence, warned data scientists against the rush to "big data" systems focused on mining massive data sets for correlations, inferences, and patterns on which to base predictions. Although Etzioni acknowledged the near-term benefits of such efforts in fields such as speech recognition and computer vision, he noted it also has major drawbacks. Etzioni said big data systems can tell you a lot of facts and even identify patterns in those facts, but they cannot reason the way humans do and he worries the attention being paid to big data will draw attention away from efforts such as those at the Allen Institute to create computer systems that can reason.

By contrast, Microsoft Research's Eric Horvitz gave a keynote on Tuesday that was brimming with enthusiasm for big data. Horvitz said he understands Etzioni's points about the limitations of big data, but believes the two approaches to data science can be complementary. He said they can move forward together and have "a huge impact in so many fields, in the short term, along the way to reasoning systems."

From The New York Times
View Full Article – May Require Free Registration


Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account