NYU Data Science newsletter – October 18, 2016

GROUP CURATION: N/A

AI is here, and it’s changing the way that we interact with technology on a daily basis. Most of the time, we don’t even realize when it’s happening. As designers, it’s going to become increasingly important to not only understand the fundamentals of AI, but also to adapt our process and mindsets accordingly. Below are six important changes that designers will need to consider.

1. First and foremost, it’s time to learn about what AI is.

Some people fear the singularity. I have a more mundane concern: are we approaching the end of intellectual property?

I’m not talking about software patents, whose disappearance I would welcome. I mean intellectual property in the broadest sense?—?the protectable economic value of our intellectual output.

You probably chafe a bit every time you learn that a certain doctor or hospital isn’t part of your insurance network. Narrowing the scope of your network helps insurers save money. They can drive hard bargains with doctors and hospitals to get lower prices and walk away from higher-priced ones.

Increasingly, insurers are offering narrow network plans. Would you enroll in one? So long as quality doesn’t suffer, consumers should welcome the lower premiums they may offer.

Researchers at the Leonard Davis Institute at Penn analyzed the relationship between network size and premiums for plans offered in the Affordable Care Act marketplaces. Plans with very narrow networks (covering care by less than 10 percent of physicians) charged 6.7 percent lower premiums than plans with much broader networks (covering care by up to 60 percent of physicians). This translates into an annual savings for an individual of between $212 and $339, depending on age and family size. For a young family of four, the savings could reach nearly $700 per year.

“The important thing is not to stop questioning,” Albert Einstein once said. “Curiosity has its own reason for existing.” In this series, we turn to NYU faculty—specialists in their fields—to address the general questions that we all encounter as we move through the world. Here, Patrick J. Egan, associate professor of politics and public policy, offers a campaign season guide to making sense of election polls.

At their core, data tell stories. They reveal patterns, show changes over time, and confirm or challenge our theories. And in cities across the country, mayors, police chiefs, and other local leaders are turning to data to help them understand and address gun violence, one of the most persistent crises they face.

Innovative, data-driven programs are showing encouraging results. To keep high school students on the right track, the city of Chicago scaled up a school-based program called Becoming a Man for seventh through tenth graders living in neighborhoods with high rates of violence. The students reflect on their life goals, observe how their automatic responses inside school and outside school differ, and learn to slow down and react more thoughtfully to these sometimes divergent social environments. An adaptive behavior on the street, like fighting back to develop a reputation of toughness that could deter future victimization, will be maladaptive in other social situations. To test the impact of the program, the University of Chicago Crime Lab built a rigorous evaluation into its rollout. After two years, they were able to show that participants were 50 percent less likely to be arrested for a violent crime than students in a control group, and those students graduated at a rate 19 percent higher than those who did not participate. This close analysis of the program affords new insight into what makes the program work, and how to enhance it and apply it in other settings.

Rather than leading to the violent downfall of humankind, artificial intelligence is helping people around the world do their jobs, including doctors who diagnose sepsis in patients and scientists who track endangered animals in the wild, experts said Thursday (Oct. 13) at the White House Frontiers Conference in Pittsburgh.

How do all of these economic decisions relate to data science? By studying the data generating process, an economist is better prepared to know if 1) the data is accurate 2) the data is complete and 3) the data is appropriate.

Machine learning and artificial intelligence have the ability to transform everything from transportation to medicine, but the industry has a long way to go to get to that point.

Bloomberg Beta Partner Shivon Zilis and Allen Institute for Artificial Intelligence CEO Oren Etzioni talked AI, robotics, machine learning, voice integration, online privacy and other futuristic topics at the GeekWire Summit 2016. The very agreeable pair concurred that AI and machine learning are showing up in a lot of applications and industries already, but the development of those concepts is in only the first inning of a nine inning game.

You’ve got to give UBS analyst Steven Milunovich major props. During IBM’s earnings call on Monday, he asked whether IBM Watson, the company’s golden child, will run on rival Amazon Web Services—and he was promptly shot down.

“No. Watson runs on our cloud, and our technology will run on IBM’s cloud,” IBM chief financial officer Martin Schroeter responded tersely.

Apple is making a visible push in the direction of AI today by hiring Carnegie Mellon University professor Ruslan Salakhutdinov for what appears to be a newly minted position: director of AI research.

After a long hiatus, Carver Mead’s prediction of the return to analog is starting to become a reality.

“Large-scale adaptive analog systems are more robust to component degradation and failure than are more conventional systems, and they use far less power,” Mead, a Caltech professor and microelectronics pioneer, wrote in an influential Proceedings of the IEEE paper in 1990.

The UK’s security services, including GCHQ, MI5 and MI6, have been unlawfully collecting and using mass datasets of personal information for more than 10 years.

The Investigatory Powers Tribunal has ruled in a judgement published online that the bodies had been collecting data without safeguards or supervision.

“bandicoot is an open-source Python toolbox to extract more than 1442 features from standard mobile phone metadata. bandicoot makes it easy for ML researchers to load mobile phone data, to analyze and visualize them.”