Data Science newsletter – November 29, 2016

GROUP CURATION: N/A

The job market for interdisciplinary Ph.Ds. with training in technology and organizations is expected to be strong over the next decade. We anticipate that most graduates will pursue academic positions in Business Schools (where the study of technology innovation and entrepreneurship is growing), Schools of Communication (where the study of organization and technology is increasing swiftly), in Schools of Information (which are being rapidly founded across the country), and in Schools of Engineering (that are increasingly interested in the social aspects of technical work). Unlike the demand for Ph.D. tenure track faculty in many other disciplines, the number of tenure track openings in these schools exceeds the number of Ph.Ds. awarded each year, and that gap is increasing.

ScienceMatters, a Swiss startup that launched in February, is trying to pave the way to a more democratized system by offering an open-source publishing platform to every scientist who wants to share his or her observations. “We are trying to publish the same way top science publications published 50 years ago,” explains Lawrence Rajendran, founder and CEO of ScienceMatters. “They used to publish exact observations, but now, competition for space is extremely high so there needs to be that wow factor.”

In other words, scientists must not only present outstanding and unique results, but they also need to craft them within an appealing narrative that pleases the editors. Therefore what drives scientists today is no longer the curiosity of discovering something new, but rather the glorification of a high-impact factor (i.e. essentially an indicator of the number of times articles published in the journal are cited).

Lighter wallets have benefits in the heavily crowded retail loyalty and rewards program business, according to Ameer Jumabhoy, MBA ’17, and his father, Asad Jumabhoy, SM ’86.

The two recently launched a Singapore-based business, UTU, which will make loyalty and reward points programs available internationally, allowing consumers to take advantage of real time usage of those points anywhere. UTU, which translates to “rewards” in Maori, launches in Thailand this month and will continue its rollout through Asia in 2016. The company plans to make its way to Europe and the United States in 2017.

“Seismiologists just don’t have enough information about the processes that are going on in Earth,” says Peggy Hellweg, Operations Manager at UC Berkeley’s Seismology Lab. “We know the stress is building up, but we don’t know the details of each particular location, and when the stress is going to be too much for it. That’s the problem with earthquake prediction.”

With strong support from Congress, President Obama has made precision medicine a focal point of the end of his term. Even with a change coming in the White House, precision medicine will remain a national priority because of its many potential benefits. The key to its success will be clinical implementation. As genetic screening tests become increasingly common, precision medicine could change the practice of medicine as we know it.

Implementing precision medicine will require health systems to decide what precision medicine will look like for them. Some systems are taking a three-pronged approach and incorporating basic science discovery, testing in clinical trials, and clinical implementation of precision medicine across separate institutes or departments, while others are combining initiatives within related areas such as biomedical research informatics and genomics. Some progressive organizations are carving out entities for precision medicine with independent budgets, staff, and decision-making abilities. There is no standard approach.

For most of its 146 years, Deutsche Bank was the embodiment of German values: reliable and safe. Now, the once-proud institution is facing the abyss. SPIEGEL tells the story of how Deutsche’s 1990s rush to join the world banking elite paved the way for its own downfall.

We consider a model of repeated online auctions in which an ad with an uncertain click-through rate faces a random distribution of competing bids in each auction and there is discounting of payoffs. We formulate the optimal solution to this explore/exploit problem as a dynamic programming problem and show that efficiency is maximized by making a bid for each advertiser equal to the advertiser’s expected value for the advertising opportunity plus a term proportional to the variance in this value divided by the number of impressions the advertiser has received thus far. We then use this result to illustrate that the value of incorporating active exploration in an auction environment is exceedingly small.

In the wake of the election, it’s clear American society is fractured. Negative emotions are running amok, and countless words of anger and frustration have been spilled. If you were to analyze this news outlet for the ratio of positive emotional words to negative ones, would you find a dip linked to the events of the past few weeks?

It’s possible, suggests a study published last week in Proceedings of the National Academy of Sciences. Analyzing Google Books and The New York Times’s archives from the last 200 years, the researchers examined a curious phenomenon known as “positive linguistic bias,” which refers to people’s tendency to use more positive words than negative words. Though the bias is robust — and found consistently across cultures and languages — social scientists are at odds about what causes it.

Modern military operations are dynamic and complex—requiring, for example, that infantry squads carry out their missions simultaneously in the 3-dimensional physical world, the cyber domain, and across the electromagnetic spectrum. As artificial intelligence becomes more advanced, the future of kinetic, cyber, and electronic warfare envisions humans and intelligent machines working together as a team. A challenge in designing human-machine systems, however, is determining how best to meld human cognitive strengths and the unique capabilities of smart machines to create intelligent teams adaptive to rapidly changing circumstances.

To address this challenge, DARPA today announced the Agile Teams (A-Teams) program, which sets out to discover, test, and demonstrate predictive and generalizable mathematical methods to enable optimized design of agile hybrid teams. A-Teams seeks to fundamentally challenge the current paradigm of human-intelligent machine systems design by changing the focus from simply using machines for automation and substitution of human capacity to an integrated fabric enabling superior collective problem solving.

Given a still image, CSAIL deep-learning system generates videos that predict what will happen next in a scene.

In 1989, the Berlin Wall began to fall, the World Wide Web made its debut, Madonna’s “Like a Prayer” topped the charts, and in Pittsburgh, a retrofitted Army ambulance called ALVINN was driving around Carnegie Mellon University without any human intervention.

Self-driving cars may seem like a very recent technological phenomenon, but researchers and engineers have been building vehicles that can drive themselves for over three decades. Research on computer controlled vehicles began at Carnegie Mellon in 1984 and production of the first vehicle, Navlab 1, began in 1986. ALVINN, which stands for Autonomous Land Vehicle In a Neural Network, was used as a test vehicle well into the 1990s.

This proto-driverless vehicle came up recently in a Twitter discussion between two engineers: Oliver Cameron, who heads an open-source self-driving car project at Udacity, and Dean Pomerleau, a CMU professor who ran the self-driving car project that gave birth to ALVINN. Cameron tweeted a video shared by some of his students of a car steering itself autonomously using only a camera.

Japan is reportedly eyeing a return to the top of the supercomputer ranks. The county’s Ministry of Economy, Trade and Industry plans to spend 19.5 billion yen ($173 million) on a new supercomputer, according to budget filings reported by Reuters, with aims of developing a machine capable of 130 petaflops.

That number would put the company in the top spot, moving the AI Bridging Cloud Infrastructure ahead of Sunway Taihulight, the supercomputer unveiled by China over the summer capable of 93 petaflops. The Ministry already has some fairly big plans for the previously unannounced supercomputer, utilizing its record-breaking speeds to help the country develop advances in AI technologies like deep learning.

GE Healthcare is teaming up with Boston Children’s Hospital to develop digital tools for the diagnosis and treatment of childhood diseases with an initial focus on diagnostic imaging for brain disease.

In this first project, GE will contribute its software and cloud computing prowess, while Boston Children’s radiologists will pitch in their clinical knowledge. The goal will be to create a decision-support tool to help radiologists interpret pediatric MRI scans of the brain, according to a statement. This has the potential to save money, time and stress associated with misdiagnosis and unnecessary testing.

Q: Tell us about your team…

David Rosenberg: I’m on the data science team in the Office of the CTO. One of our primary responsibilities is to help set the technical direction of Bloomberg as it relates to data science. This involves coordination, consulting, running projects, and both internal and external communication.

Supercomputers used to be a market unto themselves. In their time, giants such as IBM, Control Data Corporation (CDC), Evans & Sutherland (E&S), Silicon Graphics and Cray Computing ruled the supercomputing market. But with Moore’s Law restricted to only scaling transistor count, scale-out has taken over a conference series that had epitomized scale-up. Because the nature of supercomputing has changed so much in the past two decades, the market has expanded into a much broader high-performance computing (HPC) market. At this year’s SC16 conference, SGI made news by being purchased by Hewlett Packard Enterprise (HPE) and the Cray booth looked a bit forlorn. For the past few years compute silicon vendors have increased their influence on the supercomputing market. This year at SC16, most of the attention was on NVIDIA, Intel and Xilinx. Cloud infrastructure manufacturers, such as Dell EMC, HPE, IBM, Huawei, QCT, Inspur and Sugon and storage vendors, such as DDN, Dell EMC, and HPE, were also in the spotlight.

Scientists working at NYU Abu Dhabi have developed a sophisticated new theoretical model that may lead to timelier, accurate forecasts of the central Pacific El Niño (CP El Niño), an important weather-maker.

Any El Niño is a period of warmer-than-normal sea-surface temperature in the equatorial Pacific. These anomalies, and the related air pressure fluctuations, can have teleconnections – long-distance effects – on weather far away from the region. CP El Niño events can last for years, and have been “very frequent in the last 25 years,” said Andrew Majda, professor of mathematics and atmosphere/ocean science and principal investigator at NYUAD’s Center for Prototype Climate Modeling (CPCM).

Majda and Nan Chen, postdoctoral researcher at NYU’s Courant Institute of Mathematical Sciences, recently published a scholarly paper explaining what they have learned about the “genuine puzzle” of how a CP El Niño occurs.

My ECEP colleagues at the University of Massachusetts Amherst, Rick Adrion and Renee Fall, led a successful NSF alliance called CAITE. One of CAITE’s most successful strategies to improve diversity at university-level CS was to make it easier for students to transfer from community colleges. Community colleges are much more diverse.

The latest reports from Google tell us more about the obstacles that CS students still face in moving from community colleges to bachelor’s degrees, and how to make it easier.

We’ve compiled some of what we’ve heard and seen work in creating a more diverse, inclusive, and equitable science and technology workforce.

It’s early days for Big Data in China, but economists think it could reshape the way financial markets view the world’s second biggest economy. Bloomberg’s Jeff Kearns reports on “Bloomberg Daybreak: Asia.”

Launch an unmanned SBXC sailplane into the air with a catapult and it glides lazily down again in about three minutes. But when the sailplane is equipped with software to find plumes of warm air rising from the ground, known as thermals, it can ride them upward and extend its flight to a record-breaking five hours.

The software, known as ALOFT, for Autonomous Locator of Thermals, was developed by Dr. Dan Edwards and colleagues at the Naval Research Laboratory. The team tested and launched more than 20 flights with the software at the Phillips Army Airfield in Maryland last October. In total, those flights lasted more than 30 hours, proving that using algorithms to update data on thermals can help unmanned sailplanes fly much longer than previously possible.

The Isaacus initiative is preparing to collect health status and health data from across all relevant databases in Finland—not only those from the provider sector of Finnish healthcare, but also from social welfare agencies and other sources—in order to improve the capabilities of researchers to uncover social welfare and health status patterns across Finnish society, and therefore support the development of new policies and programs to address issues such as chronic disease and the social determinants of health in that country.

But mention the name Jürgen Schmidhuber in an automated quinoa lunch spot frequented by coders in San Francisco, and you are likely to get blank stares.

On a recent train ride to Zurich, Dr. Schmidhuber, an athletic 53-year-old who is co-director of the Dalle Molle Institute for Artificial Intelligence Research here, reflected on how he believed his early research was often overlooked or ignored. “It’s like much of the rest of society,” he said. “Sometimes it’s postfactual.”

Dr. Schmidhuber’s complaints are well known within the fraternity of researchers who have turned what until a half-decade ago was an academic backwater into a multibillion-dollar industry. He has been accused of taking credit for other people’s research and even using multiple aliases on Wikipedia to make it look as if people are agreeing with his posts.

Thirty years ago, a young computer scientist from McGill University, Yoshua Bengio, chose to do his mastery on a subject considered esoteric, neural networks and the ability of machines to learn for themselves. Two decades later, he will train a student who will in turn become a star in this field, Hugo Larochelle.

The computer scientist Cynthia Dwork takes abstract concepts like privacy and fairness and adapts them into machine code for the algorithmic age.

Numerical Optimization is one of the central techniques in Machine Learning. For many problems it is hard to figure out the best solution directly, but it is relatively easy to set up a loss function that measures how good a solution is – and then minimize the parameters of that function to find the solution.

I ended up writing a bunch of numerical optimization routines back when I was first trying to learn javascript. Since I had all this code lying around anyway, I thought that it might be fun to provide some interactive visualizations of how these algorithms work.

OSMDeepOD is a project about object detection from aerial imagery using open data from OpenStreetMap (OSM). The project uses the open source software library TensorFlow, with a retrained Inception V3 neuronal network.

The source of R an package is a highly structured set of files that reside in a dedicated directory on your computer. It can be beneficial to also make this directory an RStudio Project and a Git repository and, eventually, associate it with a GitHub remote. devtools ensures that your initial set up is correct and helps you keep it that way as your package evolves.

Twitter is among the commonest sources of data employed in social media research mainly because of its convenient APIs to collect tweets. However, most researchers do not have access to the expensive Firehose and Twitter Historical Archive, and they must rely on data collected with free APIs whose representativeness has been questioned. In 2010 the Library of Congress announced an agreement with Twitter to provide researchers access to the whole Twitter Archive. However, such a task proved to be daunting and, at the moment of this writing, no researcher has had the opportunity to access such materials. Still, there have been experiences that proved that smaller searchable archives are feasible and, therefore, amenable for academics to build with relatively little resources. In this paper I describe my efforts to build one of such archives, covering the first three years of Twitter (actually from March 2006 to July 2009) and containing 1.48 billion tweets. If you carefully follow my directions you may have your very own little Twitter Historical Archive and you may forget about paying for historical tweets. Please note that to achieve that you should be proficient in some programming language, knowable about Twitter APIs, and have some basic knowledge on ElasticSearch; moreover, you may very well get disappointed by the quality of the contents of the final dataset.

Swift has truly been released into the open: the conceptualization, discussion, and development of new language features, direction decisions, and changes to the features all take place on open mailing lists, with the wider community of Swift users. This is important, because it means that the direction of the language is in the hands of users and not the exclusive domain of a central planning group.

In this report, we’re going to look at Swift 3, released in September 2016, and the first significant release from the open source Swift project. Specifically, we’re going to look at three facets of Swift 3 programming and the ecosystem around its use:

To extend the range of data types supported by community repositories, we have built a prototype Image Data Resource (IDR) that collects and integrates imaging data acquired using many different imaging modalities including high-content screening, super-resolution microscopy, time-lapse imaging and digital pathology, and links them in a single resource. IDR links experimental perturbations to public genetic or chemical databases, and cell and tissue phenotypes to controlled vocabularies expressed as ontologies. By integrating the phenotypic and genetic metadata from multiple studies, IDR makes it possible to reveal novel functional networks of genetic interactions linked to specific cell phenotypes.

When people talk about their data infrastructure, they tend to focus on the technologies: Hadoop, Scalding, Impala, and the like. However, we’ve found that just as important as the technologies themselves are the principles that guide their use. We’d like to share our experience with one such principle that we’ve found particularly useful: reproducibility.

We’ll talk about our motivation for focusing on reproducibility, how we’re using Jupyter Notebooks as our core tool, and the workflow we’ve developed around Jupyter to operationalize our approach.