NYU Data Science newsletter – December 2, 2015

GROUP CURATION: N/A

U.S. intelligence agencies are in the business of predicting the future, but no one has systematically evaluated the accuracy of those predictions—until now. The intelligence community’s cutting-edge research and development agency uses a handful of predictive analytics programs to measure and improve the ability to forecast major events, including political upheavals, disease outbreaks, insider threats and cyber attacks.

The Office for Anticipating Surprise at the Intelligence Advanced Research Projects Activity (IARPA) is a place where crystal balls come in the form of software, tournaments and throngs of people.

Since the 2009 fatal shootings of 13 people at Fort Hood by a U.S. Army major and psychiatrist and the leaks of some 750,000 classified and sensitive military documents to WikiLeaks by another soldier, the U.S. Defense Department has sought technology to give analysts an advantage in finding insider threats.

The need spread, and now federal agencies employ advanced analytics and cybersecurity solutions to protect against an ever-morphing landscape of breaches, from those outside firewalls to rogue or careless employees. One of those solutions is a product called Carbon.

Benjamin Jakubowski on why he choose to study at the NYU’s Data Science MS program

… The next big platform push from a range of companies will be around machine learning. Tuning and optimizing the hardware and software stack, cobbling it together with other core open source pieces (including Spark) and rolling it out in a nice, neat vertically integrated package. We are seeing this happening in the open source machine learning software stack for certain with companies like H2O and others trying to meld together disparate pieces to form a cogent, generally applicable foundation for a broad range of machine learning tasks, and now the big players are getting board, including one of the biggest—Big Blue.

Federal regulators may need to tailor their data security expectations when it comes to technology startups, entrepreneurs and startup executives say.

Although startups appreciate the Federal Trade Commission’s attempts to raise data security consciousness among the new tech company set, the FTC could better emphasize that a reasonable data security standard is proportional based in part on the size and resources of a business, they said.

Fifteen years ago, Keith Mercier, Global Retail Leader at IBM Watson, heard a presentation about the new way consumers would shop using a computer. Though many of his colleagues dismissed the idea, Mercier stepped to the forefront of e-commerce; he led business development and strategy for Gap’s online stores, and now finds himself at another retail forefront—cognitive computing. At IBM, Keith uses cognitive learning and natural language computing to help retailers recognize patterns and uncover insights using shopper data.

Amid universities’ efforts to promote end-of-semester sporting events, ready students for final exams, and prepare for the spring semester, recent events at the University of Missouri and others have sparked a conversation about race and ethnicity in American universities. In addition to on-campus protests and demonstrations, students, news sources, and activists have taken to social media to engage in the dialogue. While news sources have largely focused on several specific institutions, racial discrimination is not isolated to the universities in the media. Taking students’ perspectives into account is important for universities to successfully recruit and serve diverse student bodies. Social media analytics allow marketers and external affairs officials to react to recent happens and create proactive strategies for aiding all students.

Measuring the volume and sentiment of conversation is crucial for devising when and how to handle sensitive topics. Analyzing conversation from November 1st to November 21st captures the conversation that was sparked by the conflict surrounding the resignation of University of Missouri President, Tim Wolfe.

This paper addresses the general problem of reinforcement learning (RL) in partially observable environments. In 2013, our large RL recurrent neural networks (RNNs) learned from scratch to drive simulated cars from high-dimensional video input. However, real brains are more powerful in many ways. In particular, they learn a predictive model of their initially unknown environment, and somehow use it for abstract (e.g., hierarchical) planning and reasoning. Guided by algorithmic information theory, we describe RNN-based AIs (RNNAIs) designed to do the same. Such an RNNAI can be trained on never-ending sequences of tasks, some of them provided by the user, others invented by the RNNAI itself in a curious, playful fashion, to improve its RNN-based world model. Unlike our previous model-building RNN-based RL machines dating back to 1990, the RNNAI learns to actively query its model for abstract reasoning and planning and decision making, essentially “learning to think.” The basic ideas of this report can be applied to many other cases where one RNN-like system exploits the algorithmic information content of another. They are taken from a grant proposal submitted in Fall 2014, and also explain concepts such as “mirror neurons.” Experimental results will be described in separate papers.

Dr. Andreas Mueller is an Assistant Research Scientist at NYU’s Center for Data Science. We recently asked him a few questions regarding his work at CDS, and his past experience in the world of data science.

In a paper published Monday in PLOS ONE, which stems from a collaboration between the Social Media and Political Participation Lab at New York University and the Research Group in Digital Media, Networks, and Political Communication at the University of Pennsylvania, we cast empirical light on the role played during protests by peripheral participants, that is, the social media users that critical voices would identify as the “slacktivists” in the network.