Data Science newsletter – June 1, 2017

Newsletter features journalism, research papers, events, tools/software, and jobs for June 1, 2017

GROUP CURATION: N/A

 
 
Data Science News



What does it mean to ask for an “explainable” algorithm?

Princeton Center for Information Technology Policy, Freedom to Tinker blog, Ed Felten


from

One of the standard critiques of using algorithms for decision-making about people, and especially for consequential decisions about access to housing, credit, education, and so on, is that the algorithms don’t provide an “explanation” for their results or the results aren’t “interpretable.” This is a serious issue, but discussions of it are often frustrating. The reason, I think, is that different people mean different things when they ask for an explanation of an algorithm’s results.

Before unpacking the different flavors of explainability, let’s stop for a moment to consider that the alternative to algorithmic decisionmaking is human decisionmaking. And let’s consider the drawbacks of relying on the human brain, a mechanism that is notoriously complex, difficult to understand, and prone to bias.


People of ACM – Fei-Fei Li

Association for Computer Machinery


from

Why is this an exciting time to be working in the areas of deep neural networks and artificial intelligence?

AI as a field is more than 60 years old. It started with a lofty goal of building intelligent machines. Since then, researchers in mostly academic labs and institutions have worked to lay the foundation of the AI field—the problem formulation, the evaluation metrics, the algorithms, and the important subfields as pillars of AI (e.g., robotics, computer vision, natural language processing, and machine learning). So I call this period “AI in vitro.” What’s exciting is that around 2010, our field entered a new phase that I call “AI in vivo.” We’ve now entered an age in which AI applications are changing the way computing is done in real-world scenarios, from transportation, to image processing, to healthcare, and more. Because of the advances in algorithms (such as neural network-based deep learning methods), computing (such as Moore’s law, Graphical Processing Units (GPUs), and soon Tensor Processing Units (TPUs), and the availability of data (such as ImageNet, etc.), AI applications are making a real difference. In fact, this is just the beginning. I see AI as the most important driving force of the Fourth Industrial Revolution and expect it to change industry as we know it. This is what makes this field exciting now.


Reflections on ROpenSci Unconference 2017

Microsoft, Revolution Analytics, Revolutions blog, David Smith


from

Last week I attended the ROpenSci Unconference in Los Angeles, and it was fantastic. Now in its fourth year, the ROpenSci team brought together a talented and diverse group of about 70 R developers from around the world to work on R-related projects in an intense 2-day hackathon. Not only did everyone have a lot of fun, make new connections and learn from others, but the event also advanced the ROpenSci mission of creating packages, tools and resources to support scientific endeavours using R.

During the two-day workshop, the attendees self-organized into teams of 4-8 to work on projects. There were 20 projects started at the ROpenSci conference, and all of them produced a working package.


A 2016 Review: Why Key State Polls Were Wrong About Trump

The New York Times, The Upshot blog, Nate Cohn


from

At least three key types of error have emerged as likely contributors to the pro-Clinton bias in pre-election surveys. Undecided voters broke for Mr. Trump in the final days of the race, or in the voting booth. Turnout among Mr. Trump’s supporters was somewhat higher than expected. And state polls, in particular, understated Mr. Trump’s support in the decisive Rust Belt region, in part because those surveys did not adjust for the educational composition of the electorate — a key to the 2016 race.

Some of these errors will be easier to fix than others. But all of them seem to be good news for pollsters and others who depend on political surveys.


Wearable system helps visually impaired users navigate

MIT News


from

Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed a new system that uses a 3-D camera, a belt with separately controllable vibrational motors distributed around it, and an electronically reconfigurable Braille interface to give visually impaired users more information about their environments.

The system could be used in conjunction with or as an alternative to a cane. In a paper they’re presenting this week at the International Conference on Robotics and Automation, the researchers describe the system and a series of usability studies they conducted with visually impaired volunteers.


Hexadite Uses AI to Automate Routine Security Incident Response

The New Stack, Susan Hall


from

The focus behind Hexadite’s security automation and orchestration (SAO) system is to tackle the security “alert fatigue.” It’s software designed to fully automate the investigation and remediation tasks typically handled by Tier 1 and Tier 2 security analysts.

And it goes against the tide of products aiming to prioritize alerts, thus reducing the number of alerts to investigate. It takes an “investigate everything” approach.

In training security analysts around the world, the founders realized that analysts spend 75 to 80 percent of their time responding to commodity malware, taking actions that could quickly be handled with an automated system, CEO Eran Barak told BetaNews.


New Tech Promises Easier Cervical Cancer Screening

Duke University, Pratt School of Engineering


from

Duke University researchers have developed a handheld device for cervical cancer screening that promises to do away with uncomfortable speculums and high-cost colposcopes.

The “pocket colposcope” is a slender wand that can connect to many devices, including laptops or cell phones.

If widely adopted, women might even use the device to self-screen, transforming screening and cure rates in low-income countries and regions of the United States, where cervical cancer is most prevalent.


The Engineers of the Future Will Not Resemble the Engineers of the Past

IEEE Spectrum, Tekla S. Perry


from

The engineers who will invent that future, [Stanford professor James] Plummer said, “will be a different breed of people than the engineers we educated in the 20th century.” There will be fewer jobs for people in a world with more automation, he pointed out, and therefore educational systems have to focus on producing tech professionals who do what computers can’t do.

For engineering education, Plummer indicated, that means a number of things. Doctoral programs likely won’t change much, he said, other than to become more interdisciplinary. But masters-level programs, at least at brick-and-mortar schools, “will just go away,” he predicts. “Instead it will be about lifelong education and just-in-time knowledge, and that will be done online.”

And undergraduate engineering education, though it will persist, will change radically.


How AI Can Keep Accelerating After Moore’s Law

MIT Technology Review, Tom Simonite


from

Feeding data into deep learning software to train it for a particular task is much more resource intensive than running the system afterwards, but that still takes significant oomph. “Computing power is a bottleneck right now for machine learning,” says Reza Zadeh, an adjunct professor at Stanford University and founder and CEO of Matroid, a startup that helps companies use software to identify objects like cars and people in security footage and other video.

The sudden thirst for new power to drive AI comes at a time when the computing industry is adjusting to the loss of two things it has relied on for 50 years to keep chips getting more powerful. One is Moore’s Law, which forecast that the number of transistors that could be fitted into a given area of a chip would double every two years. The other is a phenomenon called Dennard scaling, which describes how the amount of power that transistors use scales down as they shrink.


AlphaGo, in context

Medium, Andrej Karpathy


from

I had a chance to talk to several people about the recent AlphaGo matches with Ke Jie and others. In particular, most of the coverage was a mix of popular science + PR so the most common questions I’ve seen were along the lines of “to what extent is AlphaGo a breakthrough?”, “How do researchers in AI see its victories?” and “what implications do the wins have?”. I thought I might as well serialize some of my thoughts into a post.


How Montreal aims to become a world centre of artificial intelligence

Montreal Gazette, Betrand Marotte


from

Hopes are high that a three-day conference starting May 24 — AI Forum — will help burnish Montreal’s reputation as one of the world’s emerging AI advanced research centres and top talent pools in the suddenly very hot tech trend.

Topics and issues on the agenda include the evolution of AI in Montreal and the transformative impact AI can have on business, industry and the economy.


Human-Level AI Is Right Around the Corner—or Hundreds of Years Away

IEEE Spectrum


from

Artificial intelligence is progressing rapidly, and its impact on our daily lives will only increase. Today, there are still many things humans can do that computers can’t. But will it always be that way? Should we worry about a future in which the capabilities of machines rival those of humans across the board? For IEEE Spectrum’s June 2017 special issue, we asked a range of technologists and visionaries to weigh in on what the future holds for AI and brainlike computing.


[1705.10974] Trends in Banking 2017 and onwards

arXiv, Quantitative Finance > General Finance; Peter Mitic


from

The changing nature of the relationship between a retail bank and its customers is examined, particularly with respect to new financial concepts, debt and regulation. The traditional image of a bank is portrayed as a physical building a classical Doric portico. This image conveys concepts of service, soundness, strength, stability and security (“five-S”). That “five-S” concept is changing, and the evidence for changes that affect customers directly is considered. A fundamental legal problem associated with those changes is highlighted: a bank is no longer solely responsible for the safeguard of customer monies. A solution to this problem is proposed: banks should be jointly liable with perpetrators of criminal activity in the event of frauds as an encouragement to recognise and mitigate fraud.


‘Data bike’ pinpoints trouble spots on central Iowa trails

Des Moines Register, McKenzie Elmer


from

One lucky intern will have the best summer job ever: cycling central Iowa’s 600 miles of paved trails on a tricked-out “data bike.”

Using a 360-degree camera that sticks out like an antenna from the lime-green electric cargo bike, and a phone app that picks up vibrations caused by imperfections in the pavement, the rider will catalog trouble spots.

It’s part of the Des Moines Metropolitan Planning Organization’s effort to measure the health of recreational trails and arm the agencies that maintain them with valuable data to aid budgeting decisions and respond more quickly to damage.


When Artificial Intelligence and Social Media Marketing Collide

The Next Web, Vivian Michaels


from

Both artificial intelligence and social media marketing are getting a lot of attention nowadays because of their huge benefits and growth potential. They are benefiting both businesses and normal people in various ways. The investment has already been growing in the artificial intelligence, and the investment is further expected to grow by around 300%, according to the prediction made by the Forrester.

Talking about the social media platforms, more than 2.5 billion people are already using various social media platforms as per the statistic. This is nearly a 1/3 population of the whole planet. A marketer has the potential to reach a large no. of potential customers from all over the world with the help of various social media platforms. The artificial intelligence (AI) is already playing a key role in various business sectors, and now it’s colliding with the social media marketing.


Experts Predict When Artificial Intelligence Will Exceed Human Performance

arXiv, MIT Technology Review


from

… when will a machine do your job better than you?

Today, we have an answer of sorts thanks to the work of Katja Grace at the Future of Humanity Institute at the University of Oxford and a few pals. To find out, these guys asked the experts. They surveyed the world’s leading researchers in artificial intelligence by asking them when they think intelligent machines will better humans in a wide range of tasks. And many of the answers are something of a surprise.

 
Events



Hollywood Data Science meetup

Meetup, Hollywood Data Science


from

Los Angeles, CA Thursday, June 8, at 6:30 p.m., Netflix Hollywood. Lightning talks about data science in Los Angeles by: Lilian Coral, Ben Welsh and Yves Bergquist. [free, rsvp required]

 
Deadlines



WRF Postdoctoral Fellowship | Details and Eligibility

Washington Research Foundation aims to bring up to 10 highly creative and dedicated postdoctoral scientists each year to research institutions in Washington state to conduct groundbreaking work in the following fields: physical and life sciences, technology, engineering and math. Deadline for applications is July 16.
 
Tools & Resources



Exploring LSTMs

Edwin Chen


from

It turns out LSTMs are a fairly simple extension to neural networks, and they’re behind a lot of the amazing achievements deep learning has made in the past few years. So I’ll try to present them as intuitively as possible – in such a way that you could have discovered them yourself.


The $1700 great Deep Learning box: Assembly, setup and benchmarks

Medium, Slav Ivanov


from

After years of using a thin client in the form of increasingly thinner MacBooks, I had gotten used to it. So when I got into Deep Learning (DL), I went straight for the brand new at the time Amazon P2 cloud servers. No upfront cost, the ability to train many models simultaneously and the general coolness of having a machine learning model out there slowly teaching itself.

However, as time passed, the AWS bills steadily grew larger, even as I switched to 10x cheaper Spot instances. Also, I didn’t find myself training more than one model at a time. Instead, I’d go to lunch/workout/etc. while the model was training, and come back later with a clear head to check on it.


Hello, world! Stan, PyMC3, and Edward

Statistical Modeling, Causal Inference, and Social Science, Bob Carpenter


from

I’m on a fact-finding mission. We (the Stan development team) have been trying to figure out whether we want to develop a more “pythonic” interface to graphical modeling in Stan. By the way, if anyone knows any guides to what people mean by “pythonic”, please let me know—I’m looking for something like Bloch’s Effective Java or Myers’s Effective C++.


What Comes After SaaS?

Medium, Hacker Noon, Noah Jessop


from

It was a crisp San Francisco morning. Louis [name changed] was feeling confident, cable car track lines whirring in the background, as he stepped into his local cafe to grab his morning coffee. Strolling down the hill to his software-as-a-service company, the bright sun matches his prospects. The company, built from a three month coding binge in Louis’ tiny south bay apartment, was becoming a real business — in sight of the mythical “unkillable” stage of $10M a year in annually recurring revenue.

Later that week (February 2016), SaaS and technology stocks plunged by more than 30%. Some companies — such as data visualization platform Tableau, saw their values cut in half. This culminated in the mighty Linkedin acquiescing to Microsoft. The next Monday had a cold chill in the air — grimacing, Louis zipped his jacket up against the cold, rushing to his office, his quiet morning routine looking like an oddly distant memory. Along with a day of fog, this chill ushered in Cloud 3.0.


Data Science Podcasts

Jon Calder


from

Below are some podcasts I listen to that relate to data science and statistics. Each of them has something slightly different to offer, so if this is an area of interest to you then I recommend you give these a try!

 
Careers


Full-time, non-tenured academic positions

Research Associate/Senior Research Associate



University of Cambridge, Department of Applied Mathematics and Theoretical Physics; Cambridge, England
Internships and other temporary positions

Temp roles on Kaggle Datasets



Google, Kaggle; Bay Area, Los Angeles/Irvine, Seattle, NYC, or Boulder, CO
Postdocs

Postdoctoral Researchers, Digital Humanities (3)



University of Helsinki; Helsinki, Finland

Leave a Comment

Your email address will not be published.