NYU Data Science newsletter – January 29, 2016

GROUP CURATION: N/A

In this episode of the O’Reilly Data Show, I spoke with Eric Colson, chief algorithms officer at Stitch Fix, and former VP of data science and engineering at Netflix. We talked about building and deploying mission-critical, human-in-the-loop systems for consumer Internet companies. Knowing that many companies are grappling with incorporating data science, I also asked Colson to share his experiences building, managing, and nurturing, large data science teams at both Netflix and Stitch Fix. [audio, 35:30]

Surging costs for super-fast data feeds that traders rely on to buy and sell stocks are straining the country’s $22 trillion equity market and could ultimately hurt liquidity, according to a report from Tabb Group LLC.

Firms seeking to buy a full set of feeds from U.S. exchanges so they can trade stocks for clients, make markets and run an alternative trading system pay $816,000 per year for data, the research firm and consultant said.

Deep learning has been a popular research area due to major successes in perception tasks such as speech recognition and object classification. In the first part of my talk, I will give a brief overview of the main concepts of deep learning. I will focus on recent advances on a topic that my group has been especially actively working on: natural language processing and understanding using Recurrent Neural Networks (RNNs). In the past year, RNNs have done exceptionally well at learning to decode sequences of symbols from input signals. In the main part of this talk, I’ll review some recent successes on machine translation, image understanding, and beyond. I’ll finish the talk with a discussion of some of the next challenges for deep learning, and some exciting research and applications that people in the field have started looking at.

From the Stanford Colloquium on Computer Systems Seminar Series provided by the Stanford Computer Forum.

From recording police shootings and political protests to capturing street style and celebrity selfies, ubiquitous smart-phone cameras have become a new type of media, changing how the public sees the world.

The same tools are now systematically collecting new kinds of economically valuable data — peering into hard-to-reach places, identifying emerging trends and providing a check on official numbers. Particularly useful in developing countries, this grassroots data collection may, like the phones themselves, allow formerly lagging countries to leapfrog 20th-century approaches and establish flexible, nuanced and decentralized ways of answering economic questions.

A private company has captured 2.2 billion photos of license plates in cities throughout America. It stores them in a database, tagged with the location where they were taken. And it is selling that data.

… The Arab Spring carried the promise that social media and the Internet were going to unleash a new wave of positive social change. But the past five years have shown that liberty isn’t the only end toward which these tools can be turned.

Activists were able to organize and mobilize in 2011 partly because authoritarian governments didn’t yet understand very much about how to use social media. They didn’t see the potential, says NYU professor of politics Joshua Tucker, a principle investigator at the Social Media and Political Participation Lab at New York University. “There are a lot of reasons the people in power were slow to pick up on this,” he adds. “One of the things about not have a free press is it is harder to learn what was going on in the world.”

NYU College of Dentistry basic science researchers uncover molecular switches that turn on gene expression during spine development

A new study by basic science researchers in the Department of Basic Science and Craniofacial Biology at New York University College of Dentistry (NYUCD) sought to understand how gene expression is initiated in the notochord, the evolutionary and developmental precursor of the backbone. The notochord is an axial structure that provides support and patterning signals essential for development in all chordate embryos, including humans.

A pair of studies by a team of scientists has shed new light on the nature of a particular type of DNA sequences—tandem DNA repeat arrays—that play important roles in transcription control, genome organization, and development.

This process is of interest to researchers because misregulation of tandem DNA repeats can lead to chromosome missegregation, genome instability, and a variety of diseases, including muscular dystrophy and cancer. However, because all repeats of a tandem array have the same DNA sequence, distinguishing individual repeats from each other represents a major experimental challenge. As a result, tandem DNA sequences are among the most poorly understood structures in the genome.

In these papers, published in the journal Cell Reports, New York University biologists and their colleagues examined the regulation of tandem DNA repeats in different forms of yeast—budding yeast (to study ribosomal DNA repeats) and fission yeast (to study repeats flanking the centromere, a part of a chromosome that links sister chromatids and drives chromosome segregation during cell division).

An expert in AI separates fact from hype in the wake of DeepMind’s victory over humans in the most challenging game of all