NYU Data Science newsletter – June 29, 2015

GROUP CURATION: N/A

We extend a unified and easy-to-use approach to measurement error and missing data. In our companion article, Blackwell, Honaker, and King give an intuitive overview of the new technique, along with practical suggestions and empirical applications. Here, we offer more precise technical details, more sophisticated measurement error model specifications and estimation procedures, and analyses to assess the approach’s robustness to correlated measurement errors and to errors in categorical variables. These results support using the technique to reduce bias and increase efficiency in a wide variety of empirical research.

… I think one reason for all the attention received by LaCour’s study (or, I should say, non-study) was that it’s the most extreme case of a general problem of claims being published without real evidence.

… As a co-inventor of the Federated type meeting—see this for the story—I have curiously only gone to about half of them. One of the goals of these meetings is to get diverse researchers to talk to each other. One of the obstacles is that the language is often different. Researchers often call the same abstract concept by different names. One of my favorite examples was between “breadth-first search” and “garbage collection”—see the story here.

Another deeper reason is that they may be studying concepts that are related but not identical. We will study such an example today. It connects logicians with computer architects.

Georgia Institute of Technology researchers have developed a computing system that views gameplay video from streaming services like YouTube or Twitch, analyzes the footage and then is able to create original new sections of a game.

When officials at the European Commission proposed new data-protection rules in 2012, the prospects for science looked good. Three years on, that optimism has been replaced by concern. The best that researchers can hope for from the rules now, it seems, is that they do not make things worse. For, as politicians continue to try to protect the individual in a digital world, they risk inflicting major long-term damage on the research environment.

br/>
The rules aim to update the 1995 data-protection regulations to reflect the reality of the digital age, in which information about individuals is increasingly being used as a commodity.

A pan-European law on how data could be sourced, stored and used would, in theory, be good for research. Greater harmonization could smooth the difficulties that scientists face when they try to pool analysis of genomic data and tissue samples across national borders. Such sharing could help to organize powerful trials with large numbers of participants. But it is held back at present because different European countries have their own rules on issues such as informed consent, or on how to anonymize or pseudonymize data.

The Icahn School of Medicine at Mount Sinai today announced that Joshua Harris, co-Founder of Apollo Global Management, and his wife, Marjorie, have made a $5 million gift to establish the Harris Center for Precision Wellness. The center — the first-of-its-kind at a major U.S. academic medical institution – is part of the Icahn Institute for Genomics and Multiscale Biology and will develop innovative approaches to health monitoring and wellness management by integrating emerging technologies in digital health, data science, and genomics to enable people’s health to be treated in precise, highly individualized ways.

We are on the verge of a major upheaval in the way inventory is managed. This revolution is a result of the availability of the huge amounts of real-time data that are now routinely generated on the internet and through the interconnected world of enterprise software systems and smart products. In order to make effective use of this new data and to stay competitive, managers will need to redesign their supply-chain processes.

I am talking about going beyond using traditional historical data on past sales and stockouts. It is now possible to link data generated by all product interactions (including orders, examinations, and reviews by actual and potential customers) and transactions generated by suppliers and competitors who connect via internet web sites and cloud portals. This data can be used by material-management systems to control ordering and distribution of products throughout a company’s extended supply chain. In addition, any data that is coincident with these product interactions, that is derived from the firm’s external environment, can also be accessed and linked.

On Saturday, May 16th students, family, friends, faculty and staff of the UC Berkeley School of Information gathered under the Campanile to celebrate the class of 2015, including graduates from our first Master of Information and Data Science class. AnnaLee Saxenian, Dean of the I School, greeted the graduates, family and friends, and introduced each of the speakers.

… Even with improvements in training, scale presents a problem for deep learning. The need to fully interconnect neurons, particularly in the upper layers, requires immense compute power. The first layer for an image-processing application may need to analyze a million pixels. The number of connections in the multiple layers of a deep network will be orders of magnitude greater. “There are billions and even hundreds of billions of connections that have to be processed for every image,” says Dan Cire?an, researcher at the Manno, Switzerland-based Dalle Molle Institute for Artificial Intelligence Research (IDSIA). Training such a large network requires quadrillions of floating-point operations, he adds.

Five thousand pencil-shaped robots, densely nested in a metal hive, whir to life with a precise, dizzying choreography. Small U-shaped heads swivel into a new arrangement in a matter of seconds.

This preprogrammed routine will play out about four times per hour every night at the Dark Energy Spectroscopic Instrument. The robots of DESI will be used to produce a 3-D map of one-third of the sky. This will help DESI fulfill its primary mission of investigating dark energy, a mysterious force thought to be causing the acceleration of the expansion of the universe.

The Jefferson Project at Lake George, a joint research collaboration involving Rensselaer Polytechnic Institute, IBM Research, and the FUND for Lake George, is focused on protecting the lake and helping address the world’s looming freshwater supply challenges. … The Jefferson Project, named after Thomas Jefferson, who admired the lake, was launched 1 ½ years ago. But now that we have dozens of sensors deployed and powerful computer systems set up, we are beginning to analyze rich streams of data and to use that data to refine our multiple computer models of the workings of the lake and its watershed.

… Microsoft Research announced last summer that Microsoft Academic Search was evolving from a research project into full-scale production powered by Bing (see Making Cortana the Researcher’s Dream Assistant). In addition to integrating scholarly publications directly into Bing search results and Cortana’s notification system, we are taking full advantage of Bing’s capacity to crawl the web and generate structured information from unstructured text. Our Academic Graph of research publications, authors, journals, conferences, universities and fields of study has grown significantly, more than doubling the number of publication records of the previous iteration and offering nearly three times the number of citations between publications.

While our graph continues to grow, today we are announcing the release of a snapshot of this graph for the research community, in an effort to jumpstart new avenues of research at web scale.