NYU Data Science newsletter – June 27, 2016

GROUP CURATION: N/A

Google, Inc., isn’t just the world’s biggest purveyor of information; it is also the world’s biggest censor. … If Google were just another mom-and-pop shop with a sign saying “we reserve the right to refuse service to anyone,” that would be one thing. But as the golden gateway to all knowledge, Google has rapidly become an essential in people’s lives – nearly as essential as air or water. We don’t let public utilities make arbitrary and secretive decisions about denying people services; we shouldn’t let Google do so either.

In a special hour-long broadcast this Thursday, The Takeaway will visit Hudson Yards to examine what’s behind the “smart city” movement, which aims to improve city life through technology and big data.

HOW HAS ARTIFICIAL intelligence, associated with hubris and disappointment since its earliest days, suddenly become the hottest field in technology? The term was coined in a research proposal written in 1956 which suggested that significant progress could be made in getting machines to “solve the kinds of problems now reserved for humans…if a carefully selected group of scientists work on it together for a summer”. That proved to be wildly overoptimistic, to say the least, and despite occasional bursts of progress, AI became known for promising much more than it could deliver. Researchers mostly ended up avoiding the term, preferring to talk instead about “expert systems” or “neural networks”. The rehabilitation of “AI”, and the current excitement about the field, can be traced back to 2012 and an online contest called the ImageNet Challenge.

One of the curious features of network science is that the same networks underlie entirely different phenomena. As a result, these phenomena have deep similarities that are far from obvious at first glance. Good examples include the spread of disease, the size of forest fires, and even the distribution of earthquake magnitude, which all follow a similar pattern. This is a direct result of their sharing the same network structure.

So it’s usually no surprise that the same “laws” emerge when physicists find the same networks underlying other phenomena. Exactly this has happened repeatedly in the social sciences. Network science now allows social scientists to model societies, to study the way ideas, gossip, fashions, and so on flow through society—and even to study how this influences opinion.

To do this they’ve used the tools developed to study other disciplines. That’s why the new field of computational social science has become so powerful so quickly.

Incredible advances are being made in image processing techniques and tools, but the researchers who use them typically don’t have the opportunity to communicate with others who work on similar problems in different domains. ImageXD was founded by Berkeley and Washington partners from the Moore-Sloan Data Science Environments (MSDSE) to address these challenges.

At the inaugural ImageXD event, held at UC Berkeley’s Institute for Data Science (BIDS) on June 7-9, 2016, we gathered 50 researchers from 14 institutions representing expertise in computer vision, microscopy, materials imaging, photography, earth science, neuroscience, astronomy, software development, and more.

Experts warn that “the substitution of machinery for human labour” may “render the population redundant”. They worry that “the discovery of this mighty power” has come “before we knew how to employ it rightly”. Such fears are expressed today by those who worry that advances in artificial intelligence (AI) could destroy millions of jobs and pose a “Terminator”-style threat to humanity. But these are in fact the words of commentators discussing mechanisation and steam power two centuries ago. Back then the controversy over the dangers posed by machines was known as the “machinery question”. Now a very similar debate is under way.

More perspectives on AI:

From not working to neural networking – The artificial-intelligence boom is based on an old idea, but with a modern twist (June 25, The Economist)

Enter the Matrix: Developing a Big Red Button for AI and Robots (June 26, Medium, Mark Riedl)

Microsoft CEO Satya Nadella: Humans and A.I. can work together to solve society’s challenges (June 28, Slate)

ACCORDING to some prominent voices in the tech world, artificial intelligence presents a looming existential threat to humanity: Warnings by luminaries like Elon Musk and Nick Bostrom about “the singularity” — when machines become smarter than humans — have attracted millions of dollars and spawned a multitude of conferences.

But this hand-wringing is a distraction from the very real problems with artificial intelligence today, which may already be exacerbating inequality in the workplace, at home and in our legal and judicial systems. Sexism, racism and other forms of discrimination are being built into the machine-learning algorithms that underlie the technology behind many “intelligent” systems that shape how we are categorized and advertised to.

NYU Stern School of Business is expanding its MBA program to prepare students for future jobs in the fintech, as the sector faces a tsunami of innovation. Starting Fall 2016, NYU Stern MBAs will be able to choose a dedicated fintech specialization, which NYU Stern says is a first among top business schools, plus eight new electives.

More FinTech:

The new fintech generation: don’t fight the banks, embrace them (June 21, Lending Times)

Rogue Machine Intelligence and A New Kind of Hedge Fund (June 21, Medium, Numerai)

Consumer tech giant Apple, which has spent considerable effort positioning its products as health and fitness helpers, has just hired someone who knows Big Pharma and Big Data. Stephen Friend, a veteran of drug R&D and, more recently, a nonprofit effort to foster more collaborative biomedical research and more data sharing, is joining Apple (NYSE: AAPL) in an unspecified capacity.

The news emerged today from Sage Bionetworks, the Seattle nonprofit that Friend founded after leaving drug giant Merck (NYSE: MRK), where he was a senior research executive for eight years. Friend joined Merck in 2001 when it bought his Seattle biotech Rosetta Inpharmatics, which used genetic analysis in drug research and development.

From the passenger seat, Scott Lindstrom, manager of driver-assist technologies at Ford Motor Co. F -3.39% , assures me I’m not the first person to scream during this demo. I ask how he simulates accidents day after day. “It is very nerve-racking,” he says. Also, he’s hit the decoy plenty of times. In 2012 he even did it in front of Ford’s board of directors.

Back then the idea of self-driving cars looked, to Ford’s leadership, like a frivolous Silicon Valley moonshot. Four years later things have dramatically changed. Today Ford’s vehicle lineup features more than 30 options for semiautonomous features, including the automatic brakes I tested, and the company is aggressively working on cars that fully drive themselves. By year-end the company expects to have the largest fleet of autonomous test vehicles of any automaker.

The first paper raising the concern over AI safety came out in 1994. Only in the last couple of years have we seen a concerted interest in making sure that AI and robots cannot intentionally or unintentionally harm individuals or themselves. This is due to a number of public statements of concern from Stephen Hawking, Elon Musk, and Nick Bostrom. But it is a sentiment expressed by many people who are seeing some amazing advances play out in the public press, such as AI systems driving cars, playing Atari at human-like skill levels, and publicly beating humans in the games of Jeopardy! and Go. It is not unreasonable for people?—?and scientists?—?to start asking the question:

How can we ensure that future AI and robotic systems (a) can be controlled and (b) can be prevented from intentionally or unintentionally performing behaviors that have unintended consequences?

Alap Shah wants to reinvent one of Wall Street’s favorite wheels.

Sentieo, the new firm founded by the 35-year-old former hedge fund analyst, is the latest upstart financial data platform seeking a slice of the lucrative financial information provider market long synonymous with Bloomberg’s famous terminals.

Sentieo has signed up 85 finance clients, including hedge funds and investment banks.

They pay $500 to $1,000 a month — far less than the roughly $21,000 annual cost of a Bloomberg terminal.

More people than you’ll ever know have your online passwords. The underground trade in stolen data isn’t just for hackers looking for a payday, but for passionate collectors too—individuals who build up billion-strong archives of website credentials, voting details, physical addresses, and many more pieces of personal information on people all over the world.

These reams of data, which are often pooled together from hacked sites such as LinkedIn and MySpace, could be used for breaking into accounts or provide a helpful list of contacts for spammers. But certain netizens just amass, swap, and source them like any other collectible.

They’re data hoarders.

This package is a basic Reinforcement Learning package written in LUA for Torch. It implements some simple environments and learning policies (Policy Gradient and Deep Q Learning). It also can be easily used with the OpenAI Gym package by using lutorpy (example given in the opeanaigym directory).

As we’ve worked with Jupyter notebooks one of our frustrations has been moving around large data sets to a local or hosted notebook environment. One of the most challenging datasets to move around and analyze is satellite imagery. Each individual image can be gigabytes in size. This can make downloading and analyzing imagery slow and painful.

To make life a bit easier the Timbr.io team has created a new project called Juno. Using Juno you can connect to, execute commands on and get output from remote Jupyter kernels from within a local or hosted Jupyter Notebook, while also sending the outputs to the app for sharing with others. The remote kernels you spin up in Juno can be accessed from your notebook via our new, open source Jupyter Notebook extension called juno-magic.

We’ve finally published a neural network model which has been under development for over a year at Stanford. I’m proud to announce SPINN: the Stack-augmented Parser-Interpreter Neural Network. The project fits into what has long been the Stanford research program, mixing deep learning methods with principled approaches inspired by linguistics. It is the result of a substantial collaborative effort also involving Sam Bowman, Abhinav Rastogi, Raghav Gupta, and our advisors Christopher Manning and Christopher Potts.