NYU Data Science newsletter – September 2, 2015

GROUP CURATION: N/A

… In my opinion, there are two qualities that are most important for a successful data scientist: Firstly, it is imperative to learn how to ask the ‘right’ questions of the data and invent techniques to create action-based recommendations. Secondly, it is vital to be able to effectively communicate data-powered insights to non-technical professionals. Simply put, the formula for a data scientist might look something like this:

Data analytics + (algorithm development + machine learning) + communication skills = awesome data science!

In these waning days of summer it is common to reflect on how the days were passed. For me, it has been a wild few months. I spent these hot and sticky New York City days building the vision, team, and financing for a new company. I am incredibly excited, and honored, to report that last week was our first “official” week of work.

We are Alluvium, a team of engineers, data scientists, and executives building deeply integrated tools and services to deliver value for industries facing data challenges in the physical world. We believe that data emitted in the physical world is unruly and unrealized, but holds the keys to today’s largest business challenges.

Our mission is to build products that address these challenges, and to give the people working in these industries enhanced abilities. The journey has begun; but, how did we get here, and where are we going?

The SNLI corpus (version 1.0) is a collection of 570k human-written English sentence pairs manually labeled for balanced classification with the labels entailment, contradiction, and neutral, supporting the task of natural language inference (NLI), also known as recognizing textual entailment (RTE). We aim for it to serve both as a benchmark for evaluating representational systems for text, especially including those induced by representation learning methods, as well as a resource for developing NLP models of any kind.

Do social movements actively shape the opinions and attitudes of participants by bringing together diverse groups that subsequently influence one another? Ethnographic studies of the 2013 Gezi uprising seem to answer “yes,” pointing to solidarity among groups that were traditionally indifferent, or even hostile, to one another. We argue that two mechanisms with differing implications may generate this observed outcome: “influence” (change in attitude caused by interacting with other participants); and “selection” (individuals who participated in the movement were generally more supportive of other groups beforehand). We tease out the relative importance of these mechanisms by constructing a panel of over 30,000 Twitter users and analyzing their support for the main Turkish opposition parties before, during, and after the movement. We find that although individuals changed in significant ways, becoming in general more supportive of the other opposition parties, those who participated in the movement were also significantly more supportive of the other parties all along. These findings suggest that both mechanisms were important, but that selection dominated. In addition to our substantive findings, our paper also makes a methodological contribution that we believe could be useful to studies of social movements and mass opinion change more generally. In contrast with traditional panel studies, which must be designed and implemented prior to the event of interest, our method relies on ex post panel construction, and hence can be used to study unanticipated or otherwise inaccessible events. We conclude that despite the well known limitations of social media, their “always on” nature and their widespread availability offer an important source of public opinion data.

After decades of research and development, artificial intelligence (AI) is finally becoming a part of daily life. While we may not be fully in the age of AI just yet, there’s no denying that it’s just around the corner. … Here are ten enterprise technologies that are setting the stage for the AI era to come.

1. IBM Watson

At NYC Media Lab’s upcoming Annual Summit, which takes place on Friday, September 25th at NYU’s Skirball Center for the Performing Arts and Kimmel Center, one of the elements of the day we are most excited about is the afternoon demo session. This ‘science fair’ event, which takes place from 2–5pm in the Kimmel Center, will feature a wide range of work. In all, expect 100+ demos of new media technologies and ideas from faculty and students in disciplines from design to engineering. It’s an amazing opportunity to see some of the best work from the best talent from universities across New York City, and to get a sense of shared curiosities across the campuses.

To tell the tale of every demo would be a mammoth task, but to give a sense of the kinds of projects participants will see I’ve compiled this ‘appetizer sampler’ of a dozen examples from various participating universities. We present a range of work, in no particular order, from students and from faculty. Visit the Summit home page for a full list of participating demos, and to register.

A lack of standards and best practices can hold back businesses from becoming more data-driven, says Nate Silver, founder and editor in chief for FiveThirtyEight. But they can overcome challenges.

A lot of people complain about the math in economics. Economists tend to quietly dismiss such complaints as the sour-grapes protests of literary types who lack the talent or training to hack their way through systems of equations. But it isn’t just the mathematically illiterate who grouse. New York University economist Paul Romer — hardly a lightweight when it comes to equations — recently complained about how economists use math as a tool of rhetoric instead of a tool to understand the world.

Personally, I think that what’s odd about econ isn’t that it uses lots of math — it’s the way it uses math. In most applied math disciplines — computational biology, fluid dynamics, quantitative finance — mathematical theories are always tied to the evidence. If a theory hasn’t been tested, it’s treated as pure conjecture.

You’re smart enough to know that one aspect of this matters far more than any of the others: increased capacity for Computer Science & Engineering. But please humor our colleagues by reading the whole thing.

… As I mentioned in a recent piece on Science vs. Coaching, I think sometimes we get stuck into this idea that science is the equivalent of journal articles and complex language that only experts truly understand. People only think you are a “scientific coach/person/whatever” if you follow exact evidence based methods where you take whatever is in a journal as your method of training.

Instead, I prefer to follow Sagan’s ideas.

Similar to how hunter-gatherer’s refined their tracking skills and methods through evolutionary pressure to survive so that successful ideas stuck around , while less successful ones were discarded, coaching has followed a similar path.

Many experimental results never see the light of publication day. For a large number of these, it comes down to the data being “negative”, i.e. the expected and/or wanted effect was not observed. A straightforward example might be the testing of a soil additive that is believed to help a plant grow. If the experiment outcome shows no difference between the standard soil and the soil with the additive, then the result will end up buried in the laboratory’s archive.

But is this really the best approach to scientific results?