NYU Data Science newsletter – September 19, 2016

GROUP CURATION: N/A

We introduce a convolutional recurrent neural network (CRNN) for music tagging. CRNNs take advantage of convolutional neural networks (CNNs) for local feature extraction and recurrent neural networks for temporal summarisation of the extracted features. We compare CRNN with two CNN structures that have been used for music tagging while controlling the number of parameters with respect to their performance and training time per sample. Overall, we found that CRNNs show strong performance with respect to the number of parameter and training time, indicating the effectiveness of its hybrid structure in music feature extraction and feature summarisation.

We introduce the “Energy-based Generative Adversarial Network'” model (EBGAN) which views the discriminator as an energy function that associates low energies with the regions near the data manifold and higher energies with other regions. Similar to the probabilistic GANs, a generator is trained to produce contrastive samples with minimal energies, while the discriminator is trained to assign high energies to these generated samples.

The Allen Institute for Brain Science has published the highest resolution atlas of the human brain to date in a stand-alone issue of the Journal of Comparative Neurology. This digital human brain atlas allows researchers to investigate the structural basis of human brain function.

“To understand the human brain, we need to have a detailed description of its underlying structure,” says Ed Lein, Ph.D., Investigator at the Allen Institute for Brain Science.

Diego Oppenheimer is worried that the Googles and the Facebooks will dominate the world of artificial intelligence. … Oppenheimer and his startup, Algorithmia, are doing their part in the battle against AI hegemony. Algorithmia is what Oppenheimer calls an open marketplace for algorithms—code that companies and developers can use to beef up their websites and apps—and this marketplace now includes deep learning algorithms that handle tasks like face recognition and character recognition. Whereas OpenAI shares raw AI research, Algorithmia offers working algorithms designed to slot right into new services. “Maybe the future was already invented,” Oppenheimer says. “It’s just stuck in academic papers.”

Analytics and “big data” seem to be the next frontier in a number of arenas. Data researchers can use the large, real-time data sets that are available today to facilitate scientific discovery, improve the flow of traffic, and increase energy efficiency, among many other things.

Last year, the White House appointed the first federal Chief Data Scientist. And a few months ago, the federal government released a strategy for big data research and development. Also, numerous initiatives are under way across federal agencies to both release data sets for public use and better use data to manage federal programs.

For years, GAO’s skilled technical staff has provided insights into large data sets that support our work. We have also built up our in-house science and technology expertise. Now, with the increasing use of data across both the public and private sectors, we have established our Center for Enhanced Analytics.

The next U.S. president should move quickly to appoint a nationally recognized scientist or engineer as science adviser, according to a new set of recommendations from an academic policy think tank on the presidential transition.

The incoming administration should also rapidly install a leadership team for the White House Office of Science and Technology Policy (OSTP) and ensure that the office has access and resources to help integrate science and technology (S&T) advice at the White House and across federal agencies, according to “The Vital Role of the White House Office of Science and Technology Policy in the New Administration.”

The Baker Institute for Public Policy at Rice University, in Houston, Texas, issued the recommendations Wednesday.

Perth, Australia Deadline for abstracts’ submissions is Wednesday, October 17.

“To make the benefits of hardware acceleration available to the JVM we have to deal with the complexity and challenges of bridging Java and native C++.”

“To develop effective algorithms, rapid iteration and testing is important, torch-twrl aims to make implementing and innovating fast and easy.”

Kio Stark has always talked to strangers — she believes these fleeting moments give us new ways to fall in love with the world. She shares five ways to spark a meaningful interaction with someone you’ve never met before.