NYU Data Science newsletter – October 20, 2015

GROUP CURATION: N/A

Artificial intelligence technology (or AI) has developed rapidly during the past decade, and the effects of the AI revolution are already being keenly felt in many sectors of the economy. A growing chorus of commentators, scientists, and entrepreneurs has expressed alarm regarding the increasing role that autonomous machines are playing in society, with some suggesting that government regulation may be necessary to reduce the public risks that AI will pose. Unfortunately, the unique features of AI and the manner in which AI can be developed present both practical and conceptual challenges for the legal system. These challenges must be confronted if the legal system is to positively impact the development of AI and ensure that aggrieved parties receive compensation when AI systems cause harm. This article will explore the public risks associated with AI and the competencies of government institutions in managing those risks. It concludes with a proposal for an indirect form of AI regulation based on differential tort liability.

Imagine an urn containing two white balls and one black ball. We also have a limitless supply of extra white and black balls not in the urn, as well as a fair coin. We toss the coin three times; each time the coin comes up heads, we add a white ball to the urn, and each time the coin comes up tails, we add a black ball to the urn. We toss the coin three times, so that three new balls are put into the urn. On average we expect that the urn will end up with 3.5 white balls and 2.5 black balls.

Pólya suggested a variation on this mechanism. When it’s time to add a ball to the urn, don’t toss a coin; instead, shake up the urn, choose a ball from the urn at random, and let the color of the new ball be the color of the ball you chose. That is: pick a ball at random from the urn. If it’s white, put it back, along with a new white ball; if it’s black, put it back, along with a new black ball.

… Experts have previously said, “Don’t worry, automation has always created more jobs than it has destroyed.” But I believe that this time really is different. Yes, more jobs will be created, but I think many of them will be the kinds of jobs that can also be automated—thus no net gain (maybe even a reduction) in human employment. Creating human-level artificial intelligence (HLAI) is still the goal of many AI researchers. True HLAI implies that any task a human can perform a machine will be able to perform also. And business people (in the United States and abroad) will not hesitate to substitute more manageable and lower-cost HLAI for higher-cost human workers. Will enough non-automatable jobs be created for the unskilled and not-sufficiently educated? I think not.

… Last spring, as were preparing our data and developing our approach, a critical development took place: CMS agreed to our request to make unencrypted surgeon identifiers public for the first time. The decision grew out of a 2013 legal case in which Dow Jones & Co. (owners of the Wall Street Journal) won a lawsuit overturning an old court injunction keeping the information secret. In April 2014, CMS provided us with a crosswalk that allowed us to connect the surgeon identifiers in our data to names using the NPI registry.

That enabled us to count the number of complications and procedures for each named surgeon. We then set about risk-adjusting the “raw” complication rate. Using R we generated a mixed-effects model that took into account patient information (age, gender, and comorbidities as identified by the Elixhauser Comorbidity Index), information about the complexity of the procedure and hospital and surgeon random effects.

Articles that feature in open access journals are more likely to be cited on Twitter, a major study of how research is shared has found.

Kim Holmberg, a research associate at the University of Turku in Finland, has conducted a study looking at around 4 million “altmetric events” – the sharing of research using tools like Facebook and the academic bookmarking tool Mendeley – which might shed light on the impact of scholarly work beyond the traditional measure of citations by other academics.

LSINO-US panelistsast week, more than 300 health researchers from China and the United States converged at Stanford for the ninth Sino-U.S. Symposium on Medicine in the 21st Century. At this two-day event, health experts, thought leaders and entrepreneurs, including Lloyd Minor, MD, dean of the School of Medicine, and Jerry Yang, the Taiwanese-born co-founder of Yahoo, shared their knowledge of genomics, medical apps, and other topics related to this year’s theme: Big data in health care.

Minor kicked the symposium off saying, “We have the opportunity to harness the power of genomic data and electronic medical records, and to deliver better care, more personalized care for acute illness and, perhaps even more importantly, to predict and prevent disease before it even occurs — thereby moving the focus of medicine from sick care firmly toward health care.”

Trifacta, the global leader in data wrangling software, today announced the release of Trifacta Wrangler. Trifacta Wrangler now puts intuitive, self-service data wrangling immediately into the hands of knowledge workers everywhere. With this release, anyone who works with data in Excel or visual analysis tools, such as Tableau, can use Trifacta for free to more efficiently explore and prepare diverse data for analysis. Combining decades of research in machine learning, data visualization, human-computer interaction and distributed processing, Trifacta Wrangler empowers users of all skill levels to discover, structure, clean, enrich, validate and publish their data for analysis using a connected desktop application.

Trifacta Wrangler is available to download today on Trifacta’s web site.

Michael Blanton is an astrophysicist using data science to study galaxy evolution and map the Universe. He is the director of the Sloan Digital Sky Survey, an Associate Professor in NYU’s physics department, and an Affiliated Professor at the Center for Data Science.

What initially drew you to the field of astrophysics?

I studied engineering physics in college, and became interested in astrophysics through a planetary astronomy course I took as a sophomore. What drew me to the field was how a modicum of physical knowledge, or even a small amount of data, could be used to infer the conditions in distant and otherwise unknowable places.

Nathaniel Daw, in the Princeton Neuroscience Institute and psychology, joined the faculty this fall from New York University, where he had taught since 2007. Previously a postdoctoral fellow at University College London, Daw is a graduate of Columbia University and holds a doctorate from Carnegie Mellon University.