NYU Data Science newsletter – August 26, 2015

GROUP CURATION: N/A

We at Sift Science provide fraud detection for hundreds of customers spanning many industries and use cases. To do this, we have devised a specialized modeling stack that is able to adapt to individual customers while simultaneously delivering a great out-of-box experience for new customers, achieved by mixing the output from a “global” model – trained on our entire network of data – with the output from a customer’s individualized model.

Prior to decision forests, we used a custom-built logistic regression classifier combined with highly specialized feature engineering for our global model. While logistic regression has many great attributes, it is fundamentally limited by its inability to model non-linear interactions between features. At Sift, we tend to think of our modeling stack primarily as an enabler of our feature engineering; more powerful modeling allows us to extract the most insight from our features and can even lead to new classes of features. So when in early 2015 we stopped seeing benefits from feature engineering work, it was clear to us that we needed a major upgrade to our modeling stack.

Businesses want to be more data-driven, and vendors are capitalizing on the demand with sleek, self-service tools. But it’s culture, not tools, that creates a data-driven organization.

What is deep learning? How can it help solve unstructured data challenges? And more importantly, how can it help solve my business challenge?

These are all questions that come up frequently. Deep learning is a new area of machine learning that works to improve things like computer vision and natural language processing to solve unstructured data challenges. But that description is often easier said than understood.

To provide context, we’ve highlighted some real-world examples to help you understand deep learning, or explain its benefits to others. Check out the infographic below on Deep Learning – Real World Examples.

Every team needs talented people. In data science, talented people need not only to be good at what they do individually, but also able to challenge their colleagues to create effective new solutions to very hard problems.

How do you build a data science team to attract and retain this type of world-class talent?

… In 2000, Congress required federal law enforcement agencies to report to the Department of Justice the number of people who die during an arrest or in custody. As Mother Jones has reported, Congress has passed other laws requiring the DOJ to track the use of excessive force by U.S. police officers, but consistent, quality data collection is still not happening.

Last month, the DOJ released a new tranche of data from more than 3,000 state and local law enforcement agencies. But a New York Times investigation found this data to be virtually useless for analysis. While almost all police departments track shootings, there’s great variation in how closely departments record the use of other types of force — if they even record it at all. And historically, that information has rarely been published online for the public to see.

A collaboration between Bon Appétit and ICE, IBM Chef Watson inspires home cooks everywhere to discover unexpected flavor combinations, address everyday mealtime challenges in creative ways and bring new ideas to the kitchen.

The free Watson-powered app includes knowledge gained from training Watson to understand 10,000+ recipes from the Bon Appétit database, in addition to how ingredients are used in different dishes and cooking styles.

One of my favorite things about R is that it allows me to follow up on interesting news stories. Consider this interview on EconTalk about the history of fracking in America. Russ Roberts interviewed Gregory Zuckerman about his book The Frackers. One thing that struck me were the stories of how North Dakota is being transformed by the fracking boom. North Dakota sits on the Bakken formation which, due to fracking, is now able to be monetized.

Here are two maps I made which demonstrate North Dakota’s recent demographic changes. The first shows that between 2010 and 2013 North Dakota’s Per Capita Income grew at a rate of 15%, significantly above any other US state. The second one shows that North Dakota’s Median Age decreased by 2%, significantly below any other US state. Today I will demonstrate how to create these maps in R.

While many recognize Yann LeCun as the inventor of convolutional neural networks, the momentum of which has ignited artificial intelligence at companies like Google, Facebook, and beyond, LeCun has not been strictly rooted in algorithms. Like others who have developed completely new approaches to computing, he has an extensive background in hardware, specifically, chip design, and this recognition of specialization of hardware, movement of data around complex problems, and ultimately, core performance, has proven handy.