NYU Data Science newsletter – March 25, 2016

GROUP CURATION: N/A

Sidewalk Labs is partnering with the Dept. of Transportation and Smart City Challenge finalists to create an analytics platform to improve transit and engage citizens.

The botanical value of angiosperm leaf shape and venation (“leaf architecture”) is well known, but the astounding complexity and variation of leaves have thwarted efforts to access this underused resource. This challenge is central for paleobotany because most angiosperm fossils are isolated, unidentified leaves. We here demonstrate that a computer vision algorithm trained on several thousand images of diverse cleared leaves successfully learns leaf-architectural features, then categorizes novel specimens into natural botanical groups above the species level. The system also produces heat maps to display the locations of numerous novel, informative leaf characters in a visually intuitive way. With assistance from computer vision, the systematic and paleobotanical value of leaves is ready to increase significantly.

Demand for data scientists is growing, driven by companies and government agencies that are flooded with data and struggling to make sense of it.

But what exactly do data scientists do? Wall Street Journal reporter Deborah Gage spoke with one— Ram Narasimhan, who works at General Electric Co. ’s GE Digital in San Ramon, Calif.—to shed some light on the profession.

The past quarter-century has seen a wealth of advances in neuroscience, from neuroimaging techniques that make it possible to see inside the live human brain to noninvasive electrical brain stimulation to selective activation of neurons using laser light for research in animals. The popularity of the field has exploded, with membership of the Society for Neuroscientists steadily climbing from its founding in 1969 to 40,000 members today. Yet little if any of this activity has resulted in improvements in clinical care for the mentally ill.

Preparing for social sabotage should be mandatory before social tech tools reach the general public.

Microsoft is testing a new chat bot, Tay.ai, that is aimed primarily at 18 to 24 year olds in the U.S.

Tay was built by the Microsoft Technology and Research and Bing teams as a way to conduct research on conversational understanding. The Bing team developed a similar conversational bot, Xiaolce, for the Chinese market, back in 2014. Microsoft execs dubbed Xiaolce “Cortana’s little sister.”

The White House Office of Science Technology Policy has unveiled plans to co-host four open data roundtables, with the first to get underway Thursday, as part of a continuing push to advance the use of federal data. … The roundtables, which will be co-hosted and conducted by the Center for Open Data Enterprise, which conducted similar roundtables last year, will focus on four challenges confronting the open data community.

We have Rachel Binx on the show to discuss her experience developing data visualization software for NASA JPL. … On the show we talk about the project, the process for NASA data collection and analysis, and how to write code that goes into space! [audio, 37:52]

Microsoft co-founder Paul Allen’s Vulcan Inc. investment company will develop four six-story office buildings for Google in Seattle’s South Lake Union neighborhood, a short distance from Amazon’s headquarters, the companies just announced. … Google will be moving from its current location in Seattle’s Fremont neighborhood.

In the cloud computing business, Google’s technology prowess is rarely questioned. Its commitment, however, has been doubted.

Google, which trails Amazon and Microsoft in the fast-growing market, hopes to change the industry perception that it is halfhearted about its cloud computing service with product announcements, technology demonstrations and strategy briefings at a two-day conference in San Francisco.

More news from Google/Alphabet:

Google’s Greene Hastens Cloud Expansion to Catch Amazon (Bloomberg Business, March 22)

Google announces Cloud Dataflow with Python support (Google Cloud Platform, Google Cloud Big Data Blog, March 22)

Innovation For Rent Is The Heart Of Google’s Cloud (The Next Platform, March 23)

Machine Learning in the Cloud, with TensorFlow (Google Research Blog, Slaven Bilac, March 23)

Google to move to new 4-building complex in Amazon’s backyard in Seattle, developed by Paul Allen’s Vulcan Inc. (GeekWire, March 24)

Alphabet’s Sidewalk Labs hopes smart cities will go with its Flow analytics platform (TechRepublic, Alex Howard, March 18)

We are doing something very special for the next NYU-wide data science
showcase, which will be on April 4th, 4-6pm.

There’s no talk. Rather, it will be a an “open dating” session,
specifically designed to help you meet new faculty and forge new
collaborations. This event is for faculty only, and we need
you to sign up here.

Following this event there will be a seed grant funding opportunity
which will be open to the event’s participants.
However, we want you to participate even if you are not interested in
the funding!

The deadline to register for the event is Friday, March 25. (Today!)

For nearly half a century, Artificial Intelligence (AI) has been more science fiction than science: exciting, possible, but just out of reach. And despite significant advances, “strong AI” in many ways remains elusive. Best-selling author and entrepreneur Gary Marcus [from NYU CDS] provides a cognitive scientist’s perspective on AI [in this Santa Fe Institute talk]. What have we learned? What are we still struggling with? Perhaps most compelling, is there anything programmers of AI can still learn from studying the science of human cognition? [video, 1:06:01]

In this article we introduce the main family of algorithms, known collectively as Markov Chain Monte Carlo (MCMC), that allow us to approximate the posterior distribution as calculated by Bayes’ Theorem. In particular, we consider the Metropolis Algorithm, which is easily stated and relatively straightforward to understand. It serves as a useful starting point when learning about MCMC before delving into more sophisticated algorithms such as Metropolis-Hastings, Gibbs Samplers and Hamiltonian Monte Carlo.

Once we have described how MCMC works, we will carry it out using the open-source PyMC3 library, which takes care of many of the underlying implementation details, allowing us to concentrate on Bayesian modelling.

R Tools for Visual Studio (RTVS), currently available as a Public Preview release, is a new tool from Microsoft for creating R programs using Visual Studio. RTVS is free, and Open Sourced under the MIT license. It can be downloaded by following the instructions here, and you can read our documentation here. [video, 13:10]