NYU Data Science newsletter – April 29, 2016

GROUP CURATION: N/A

Idibon, a startup that has developed technology for analyzing text in many languages, could be shutting down soon, VentureBeat has learned.

The company ran out of the money needed to keep operating independently, and acquisition talks recently fell apart, a source familiar with the matter told VentureBeat. Employees were told earlier this week that Friday would be their last day, the source said. Cofounder and CEO Rob Munro confirmed that information, but he also said the startup has received two more offers.

Traffic congestion is one of the most notable problems arising in worldwide urban areas, importantly compromising human mobility and air quality. Current technologies to sense real-time data about cities, and its open distribution for analysis, allow the advent of new approaches for improvement and control. Here, we propose an idealized model, the Microscopic Congestion Model, based on the critical phenomena arising in complex networks, that allows to analytically predict congestion hotspots in urban environments. Results on real cities’ road networks, considering, in some experiments, real-traffic data, show that the proposed model is capable of identifying susceptible junctions that might become hotspots if mobility demand increases.

We’ve been quite busy, preparing for MLconf Seattle in 3 ½ weeks, but I wanted to take a few moments to share the highlights from our 3rd MLconf NYC, which occurred on April 15th. … EEG technology seems to have a sweet spot, last year Ted Willke and Irina Rish astounded the crowd with their stories. This time, it was Jennifer Marsman’s turn with her energetic presentation and futuristic EEG device that excited the audience. Braxton McKee’s talk about how he envisions scalable platforms with more automation and help from the compilers and program analysis seemed to create discussions in the hallways. This MLconf seemed to raise a bit of controversy and discussion during the breaks.

Between academic discovery and product development lurks a lull in research funding that inventors call the “chasm of death,” where a prototype or a proof of concept can feel just out of reach.

To address that development gap, five projects initiated by faculty at Harvard John A. Paulson School of Engineering and Applied Sciences and in the Faculty of Arts and Sciences have been selected to receive funding of between $70,000 and $100,000 each through Harvard’s Physical Sciences & Engineering Accelerator.

The WIRED NextList includes: Yann LeCun & Greg Brockman among many others, saying …

The world’s most powerful tech companies are racing to create machines that think like humans. But rather than hide their advances from rivals, the brightest minds in artificial intelligence are taking a counterintuitive approach: They’re giving away their secrets.

I’m spending the year in Google’s Visiting Faculty program. … One of my explicit goals was to steal ideas from Google that I could feed back into teaching and mentoring.

“Sesame Street,” arguably the most innovative force in preschool education in history, announced Wednesday morning a new multi-year partnership with IBM’s artificial intelligence juggernaut, Watson.

Hot on the heels of a new artificial intelligence-capable thermal imaging camera, chip maker Movidius has yet another AI implementation up its sleeve: a USB stick that can allow pretty much any Linux computer to handle advanced neural networks, one of the building blocks of AI. The San Mateo, Calif.-based company announced the device, called the Fathom Neural Compute Stick, today along with its Fathom deep-learning software framework. Together, the two products will enable device manufacturers to move AI processing from the cloud to native deployment in end-user devices.

Chipotle announced its first loss as a public company Tuesday. But two weeks earlier, an unlikely source —the social media app Foursquare — had beat Chipotle to the punch, predicting in a blog post that the burrito maker’s sales would drop nearly 30 percent. Chipotle made it official Tuesday afternoon — reporting a drop of 29.7 percent.

The remarkably accurate prediction from a company consumers know for restaurant tips and the ability to check in at locations highlights the emerging power of the gobs of data our smartphones collect and the opportunity for savvy companies to convert that information into piles of cash.

This month marks the centennial of the birth of Claude Shannon, the American mathematician and electrical engineer whose groundbreaking work laid out the theoretical foundation for modern digital communications. To celebrate the occasion, we’re republishing online a memorable profile of Shannon that IEEE Spectrum ran in its April 1992 issue.

first example with gym

This is a very basic DQN (with experience replay) implementation, which uses OpenAI’s gym environment and Keras/Theano neural networks.