NYU Data Science newsletter – October 26, 2015

GROUP CURATION: N/A

Convolutional Neural Networks are great: they recognize things, places and people in your personal photos, signs, people and lights in self-driving cars, crops, forests and traffic in aerial imagery, various anomalies in medical images and all kinds of other useful things. But once in a while these powerful visual recognition models can also be warped for distraction, fun and amusement. In this fun experiment we’re going to do just that: We’ll take a powerful, 140-million-parameter state-of-the-art Convolutional Neural Network, feed it 2 million selfies from the internet, and train it to classify good selfies from bad ones. Just because it’s easy and because we can. And in the process we might learn how to take better selfies 🙂

In a 10th-floor meeting room in Crystal City, Malynda Chizek Frouard rode a red Capital Bikeshare bike to the front of the room as about 60 fellow techies and transportation buffs watched.

She had come to share her new Web site, CaBiBrags.com, which allows Capital Bikeshare users to compete for the highest mileage.

“I’m pretty sure I win the ‘Queen of the Nerds’ award tonight!” said Chizek Frouard, 31, a resident of Bethesda, pointing to the bright-blue astronaut costume she’d worn in the pre-Halloween spirit.

As part of a fireside chat hosted by Chancellor Nicholas Dirks, about 20 students gathered at University House on Wednesday to give their input on the trajectory of a pilot data science class.

The four-unit class, titled “Foundations of Data Science,” entered its preliminary planning stages in fall 2014, involving professors in fields ranging from computer science to history. The curriculum was finalized at the start of the current semester, and faculty are now using students’ experiences to determine the future of the course.

“I’ve been teaching, one way or another, for about 30 years, and I have never been so excited,” said Ani Adhikari, the principal instructor for the course, during the chat. “For the courage you have shown to take a course on spec, we are very, very grateful.”

Big data is opening up all kinds of doors in health and medicine. Genomics and other ‘omics, wearable technologies, combining detailed patient datasets for precision medicine – the possibilities seem endless.

In the most recent episode of Raw Data, a new podcast produced by Worldview Stanford, my co-producer Mike Osborne and I explored the landscape of big data in medicine. We had the opportunity to interview several Stanford researchers, including Michael Snyder, MD, who talked to us about his working taking the quantified self to the extreme; Euan Ashley, MD, who explained to us the promise that wearables hold for early diagnosis of cardiovascular disease; and Sylvia Plevritis, PhD, who discussed how patient datasets are revolutionizing cancer biology research.

The original plan for Foursquare was not about building a check-in app—it was about data.

Furthermore, creating a recommendation engine and software for navigating unfamiliar cities, based on said data, was always part of the agenda, said co-founder and CEO Dennis Crowley.

“The point was to build something that connects little bits of data every single day about where all these people are going,” he said during the Street Fight Summit.

Connectivity and integrated data are moving ahead as the best minds in metadata, digital identifiers and networked discovery collaborate to ensure that appropriate credit and recognition are given for scientific outputs of all types. Once connected to the research infrastructure through a personal and unique digital identifier, a researcher’s work becomes more discoverable and their need to reenter information into multiple systems to comply with funder, institutional and publisher requirements ultimately will be reduced.

Researchers call it sheer coincidence that a newly discovered piece of space junk is officially designated WT1190F. But the letters in the name, which form the acronym for an unprintable expression of bafflement, are an appropriate fit for an object as mysterious as it is unprecedented.

Scientists have worked out that WT1190F will plunge to Earth above the Indian Ocean on 13 November, making it one of the very few space objects whose impact can be accurately predicted. More unusual still, WT1190F was a “lost” piece of space debris orbiting far beyond the Moon, ignored and unidentified, before being glimpsed by a telescope in early October.

An observing campaign is now taking shape to follow the object as it dives through Earth’s atmosphere, says Gerhard Drolshagen, co-manager of the European Space Agency’s Near-Earth Object office.

The plummeting success rates in grant applications in the last decade are linked to flat budgets and more resubmitted proposals.

Research papers and data products are key outcomes of the science enterprise. Governmental, nongovernmental, and private foundation sponsors of research are increasingly recognizing the value of research data. As a result, most funders now require that sufficiently detailed data management plans be submitted as part of a research proposal. A data management plan (DMP) is a document that describes how you will treat your data during a project and what happens with the data after the project ends. Such plans typically cover all or portions of the data life cycle—from data discovery, collection, and organization (e.g., spreadsheets, databases), through quality assurance/quality control, documentation (e.g., data types, laboratory methods) and use of the data, to data preservation and sharing with others (e.g., data policies and dissemination approaches).

LEARN MORE ABOUT YOUR DNA AND CONTRIBUTE TO IMPORTANT SCIENTIFIC RESEARCH

We will compare your DNA with reference data from different populations to see where in the world your ancestors might have lived

In a recent InfoWorld column entitled “3 Key Ways Hadoop is Evolving”, Serdar Yegulalp discusses the ways in which the big data platform is shifting, with the scope of the article focusing on news that broke at the recently completed Strata + Hadoop World conference. According to the author, there has been a lot of change within the big data landscape specifically over the last six months. With a focus on sessions, speeches, and new technologies presented at the event, he then outlines the recent significant trends.

At Solutions Review, we have a vendor-neutral focus on the big data landscape, and with that, this post seemed like it would be worth summarizing. With all that said, here are the three key ways Hadoop is evolving, according to Yegulalp.