NYU Data Science newsletter – March 16, 2016

GROUP CURATION: N/A

Microsoft, owner of the popular video game, revealed that computer scientists and amateurs will be able to evaluate and develop AI software using its virtual landscapes from July.

The company says Minecraft is more “sophisticated” than existing AI research simulations and cheaper to use than building a robot.

Autonomous vehicles will join human drivers on our roads sooner than most people think. At Baidu, we plan to put commercial, self-driving cars on the roads by 2018. But the way to make them safe is not to make them act just like human driven cars—in fact, the limitations of today’s technology mean that’s not feasible. Instead, we should make modest changes to our infrastructure, program these cars to behave as predictably as possible, and teach the public new ways to interact with them.

These changes can accommodate autonomous cars’ weaknesses, like their inability to interpret a construction worker’s hand gestures, and take advantage of their strengths: their unending vigilance, their lack of blind spots.

What makes an idea, a brand, or a behavior catch fire? This hour, TED speakers explore the mysteries behind the many things we spread: laughter and sadness, imagination, viruses and viral ideas. [audio, 5 segments of approximately 10 minutes each]

… had a small surge, and not just from scientists whose august stature protects them from risk. On Twitter, preprint insurgents are celebrating one another’s postings and jockeying for revolutionary credibility.

The Minister of Science, Kirsty Duncan, today announced Dr. Erik Snowberg as the Canada Excellence Research Chair (CERC) in Data-Intensive Methods in Economics at The University of British Columbia (UBC). Dr. Snowberg comes to Vancouver from the California Institute of Technology. He will become Canada’s first CERC in the social sciences and the country’s 26th CERC overall.

Dr. Snowberg aims to help transform UBC into the world leader in data-driven political economy, a field that focuses on the relationship between politics and the economy. He and his research team are developing new ways to analyze big data — massive datasets of information produced and stored by companies, governments and private citizens. Dr. Snowberg’s research will lead to new analytical tools for governments to use when making policy decisions.

Mercedes-Benz offers the S-Class sedan with a growing array of options such as carbon-fiber trim, heated and cooled cupholders and four types of caps for the tire valves, and the carmaker’s robots can’t keep up.

With customization key to wooing modern consumers, the flexibility and dexterity of human workers is reclaiming space on Mercedes’s assembly lines. That bucks a trend that has given machines the upper hand over manpower since legendary U.S. railroad worker John Henry died trying to best a motorized hammer more than a century ago.

“Robots can’t deal with the degree of individualization and the many variants that we have today,” Markus Schaefer, the German automaker’s head of production, said.

About a decade ago, biologists set out to count river dolphins along a murky, meandering stretch of the upper Amazon River between Colombia and Peru. Although these distant freshwater cousins of seagoing whales face growing threats from pollution, dams, and hunting, the results seemed to suggest they were doing well. By the end of the 2007 survey, the team had spotted 116 groups of pink river dolphins (Inia geoffrensis), as well as 220 groups of a second species, the gray river dolphin or Tucuxi (Sotalia fluvia). Those numbers are higher than in two prior tallies done in 1993 and 2002.

But because the tallies had all been done differently, conservation scientists could not say for sure whether the dolphin populations were growing, merely stable, or already in dangerous decline. Now, researchers have used a 250-year-old statistical approach to draw firmer—and potentially worrisome—conclusions about the dolphin population trends.

Looking to help speed the development of 5G technology related to millimeter wave spectrum, NYU Wireless opened up a channel simulator tool based on the group’s research and experiments. NYU Wireless is a Brooklyn, New York,-based research center focused on next-generation wireless networks and devices; Ted Rappaport serves as the founding director of the institution.

As “5G” technology and standards continue to evolve, millimeter wave is being looked at to support more efficient and high-capacity data transmission. Specifically, the new software supports channel simulation in frequencies ranging from 28 GHz to 73 GHz, with French operator Orange already using the technology.

… scientists are getting more information out on social media than we used to, and it’s a very flexible, powerful way to do it,” he said. “But there are these side effects.”

How can a city’s infrastructure cope with this sudden influx of pilgrims, and the demands for food, water, shelter, and safety they bring with them? And how can communities worldwide become more resilient and livable? These kinds of questions drive Associate Professor Ramesh Raskar, a native of Nashik [India], and his Camera Culture group at the MIT Media Lab.

With support from the MIT Tata Center for Technology and Design, Raskar’s group is forging a new co-innovation model linking researchers at MIT with energetic students in India to work on problems across key fields such as health care, education, and the environment.

The phrase “big data” — and the vast, daunting numbers it implies — is thrown around a lot these days. To illustrate the concept, Helen Nissenbaum, professor of media, culture and communication and computer science at New York University, turned to something much more familiar: the emoji.

In a lecture titled “Must Privacy Give Way to Use Regulation?” Nissenbaum discussed the incongruity between consumers’ expectations and companies’ practices when it comes to data collection. To illustrate cases where business interests align with the public’s interests, Nissenbaum used two smiley emojis. Where consumers saw no benefit, she selected a neutral face; where data use was contrary to the public interest, she used a frown.

“We are given the story that we should not resist (big data) collection because we benefit so much from data machinations,” Nissenbaum said. “But the reality is that business imperatives often don’t match the public interest.”

I’m really excited to announce a preview of a new feature in Apache Kafka called Kafka Streams. Kafka Streams is a Java library for building distributed stream processing apps using Apache Kafka. It will be part of the upcoming Kafka 0.10 release and we’ve made a preview version available to make it easy to try out now. The Kafka Streams source code is available under the Apache Kafka project.

Planet OS is excited to announce the launch of our first API designed specifically for geospatial and Earth observation data. The Planet OS API provides developers, researchers, and climate specialists with a new tool for working with Earth data from the world’s most respected providers. Using the Planet OS API, it is now possible to build scalable applications and data-driven solutions without developing custom interfaces for each individual dataset.

Designing for trust is a well understood topic across the hospitality industry, but our efforts to democratize hospitality mean we have to rely on trust in an even more dramatic way. Not long ago our friends and families thought we were crazy for believing that someone would let a complete stranger stay in their home. That feeling stemmed from the fact that most of us were raised to fear strangers. … How best to support trust?—?particularly between groups of people who may not have the opportunity to interact with each other on a daily basis?—?is a core research topic for our data science and experience research teams. In preparation for [Joe Gebbia’s TED talk], we reflected on how we think about trust, and we pulled together insights from a variety of past projects.

Scikit Flow is a simplified interface for TensorFlow, to get people started on predictive analytics and data mining. It helps smooth the transition from the Scikit-learn world of one-liner machine learning into the more open world of building different shapes of ML models. You can start by using fit/predict and slide into TensorFlow APIs as you are getting comfortable. It’s Scikit-learn compatible so you can also benefit from Scikit-learn features like GridSearch and Pipeline.