NYU Data Science newsletter – September 30, 2016

GROUP CURATION: N/A

“I developed a simple gem tensorflow.rb and started with making use of SWIG.”

Franz Och, a star computer scientist known as the chief architect of Google Translate, has left his job as chief data scientist at J. Craig Venter’s Human Longevity Inc. to join Grail, the ambitious pan-cancer screening company co-founded by Illumina.

This was the first year for this new conference, and it was a blast: the energy from the 1,000 attendees was palpable. I covered Joseph Sirosh’s keynote presentation yesterday, but today I wanted to highlight a few other talks from the program now that the recordings are available to stream.

The night sky is anything but tranquil, and a new generation of robot telescopes is getting a handle on the cornucopia of fleeting flashes and flare-ups.

Harnessing the power of open data is key to developing the smart cities of the future. But not all governments have the capacity—be that funding or human capital—to collect all the necessary information and turn it into a tool. That’s where Mapbox comes in.

Mapbox offers open-source mapping platforms, and is no stranger to turning complex data into visualizations cities can use, whether it’s mapping traffic fatalities in the U.S. or the conditions of streets in Washington, D.C., during last year’s East Coast blizzard.

“The by far biggest issue I see is data science solving irrelevant problems. This is a huge waste of time and energy. The reason is typically that whoever has the problem is lacking data science understanding to even express the issue and data scientists end up solving whatever they understood might be be the problem, ultimately creating a solution that is not really helpful (and often far too complicated).”

The Administration is releasing a fact sheet of key accomplishments over the past eight years in opening government data to spur innovation, opportunity, and economic growth. The release of this list comes in conjunction with today’s White House Open Data Innovation Summit.

The National Center for Ecological Analysis and Synthesis (NCEAS) is launching an innovative “Data Task Force” to support synthesis research with a new $2.1 million grant from the Gordon and Betty Moore Foundation. In contrast to mining Big Data, synthesis research recognizes the value of small, information rich datasets generated by scientists all over the world. However, collecting, combining, and “wrangling” these many discrete data sets for synthesis presents a daunting challenge.

Microsoft announced a new group that will help accelerate this evolution. It will include Microsoft Research along with the company’s Information Platform Group, Bing and Cortana product groups, and our Ambient Computing and Robotics teams. The combined group will include more than 5,000 computer scientists and engineers.

As we all know too well, making all scientific data technically and legally accessible to img_20160913_133322all researchers is an ambitious task complicated by constantly evolving social and technical barriers. It is fair to say that we are making progresses in this direction. At Scidatacon 2016, we examined several concrete solutions that can facilitate openness of scientific data or, if you prefer, make sure data are FAIR (findable, accessible, interoperable and reusable).

However, it seems that the more we learn about how to make data open, the least we know about how exactly data will be reused by the scientific community, which means by the researchers who generated the data and should have a primary interest in accessing it.

Application Deadline: November 10. Current faculty and students at any accredited university located in New York City are eligible to apply to this business plan competition for media technologies in NYC.

GLVisualize, a library for fast and interactive graphics written in Julia.

We are excited to announce the release of YouTube-8M, a dataset of 8 million YouTube video URLs (representing over 500,000 hours of video), along with video-level labels from a diverse set of 4800 Knowledge Graph entities. This represents a significant increase in scale and diversity compared to existing video datasets. For example, Sports-1M, the largest existing labeled video dataset we are aware of, has around 1 million YouTube videos and 500 sports-specific classes–YouTube-8M represents nearly an order of magnitude increase in both number of videos and classes.

When a different analytic model is used in training versus deployment, results can be disastrous. Here’s how to avoid the anti-pattern.

A central repository of true and agreed upon definitions of key metrics means your goals will have the same definition across teams, your conversations with investors will be consistent, and your analytical cycle time will be reduced.

“Kafka is the industry-leading open source solution for delivering data-intensive applications and managing event streams, and is backed by an active open source community and ecosystem. Now combined with Heroku’s developer experience and operational expertise.”