Data Science newsletter – October 18, 2018

GROUP CURATION: N/A

“I’m one of the 90 percent of the population who are not intrinsically driven to be active. I’m a great fan of ‘mindless’ activity, where we get our physical activity as a natural part of the day – the way our forefathers lived,” [Abby King] laughs.

Dr. King and her colleagues at Stanford are attacking the problem with big data culled from health apps, as well as with local data collected by residents themselves. Their aim is to unmask the true story about our activity – or lack of it – and support practical neighborhood improvements to get us back on track.

“Many of us are pretty tired at the end of the day. If you ask people if they were active, they’ll say yes. It feels that way, but we can see from the data they really weren’t.”

Hospitals and health plans are increasingly using the huge amount of medical data they collect for research. It’s a business worth billions of dollars, and sometimes those discoveries can be the foundation of new profit-making products and companies.

When a company profits from your data, should you get a cut?

This isn’t just a hypothetical question. When Steven Petrow was 26 years old, back in 1984, he was treated for testicular cancer at Memorial Sloan Kettering Cancer Center in New York City. [audio, 4:05]

Students at the Virginia Tech Carilion School of Medicine are the first in the country to use portable ultrasound machines equipped with 12-lead EKG technology, which allows them to view images of the heart while also measuring its electrical activity through electrodes attached to the body.

“People recognize us as a leader in ultrasound teaching,” said Paul Dallas, director of the school’s ultrasound curriculum. “Our school recognized the importance of ultrasound teaching early on by building it into students’ basic science instruction starting in their first year. These new machines are an indication of our continued commitment.”

In fact, according to Dallas, the medical school is only the third in the country to have a designated ultrasound curriculum, one that is introduced within the first few weeks of school.

The Fraunhofer MEVIS Institute for Medical Image Computing – in short: Fraunhofer MEVIS – is one of the world’s leading research centers in digital medicine. In order to prepare computer science students for challenges in this area, the institute and the University of Bremen are now cooperating even more closely in teaching. This winter semester saw the introduction of a new study area called Medical Computing in the Faculty of Mathematics / Computer Science.

Whether in health care, diagnosis, surgery, or treatment: today, digital medicine plays an important role in everyday clinical life. The aim is for physicians to make the best possible use of the opportunities offered by big data, artificial intelligence, and image-based medicine. The methods in digital medicine are developed in computer science. The new study focus of Medical Computing in the Faculty of Mathematics / Computer Science at the University of Bremen reflects groundbreaking developments in this field. For example, students gain insights into medical image processing and methods that can help physicians analyze increasingly complex situations. In order to create a connection to practice, clinical staff will also come to the university and be integrated into the teaching operations.

In recent years, however, widespread access to the internet and social media has reduced physicians’ dominion over medical information and, increasingly, over patients’ access to medical products and services. It is no longer the case that patients simply visit their physicians, describe their symptoms, and wait for the differential diagnosis. Today, some patients arrive at the physician’s office having thoroughly researched their symptoms and identified possible diagnoses. Indeed, some patients who have lived with rare diseases may even know more about their conditions than some of the physicians with whom they consult.

The expanding availability of direct-to-consumer (DTC) tests and screens has further diminished physicians’ control over patients’ access to medical resources. Some tests that once required expensive equipment and an office or hospital visit can now be done by a consumer at home. For example, consumers can perform a do-it-yourself electrocardiogram on a $99 device that interfaces with a smartphone app. The use of DTC laboratory tests is also increasing. These tests run the gamut of scientific legitimacy. Some, such as an at-home blood test for food sensitivity, do not meet standards of clinical validity.3 Others, by contrast, are considered high quality and, until recently, would have required a physician’s order. For example, depending on the state, consumers can order an array of laboratory tests, including complete blood cell counts, comprehensive metabolic panels, hepatitis C screening, and a variety of sexually transmitted disease screening panels.

IBM (NYSE: IBM) today introduced AI OpenScale, a new technology platform that addresses key challenges of artificial intelligence (AI) adoption, such as concerns over how AI applications make decisions, the global shortage of AI skills and the complexities of working with disparate AI tools from multiple vendors.

IBM’s new technology platform is the first of its kind. It will enable companies to manage AI transparently throughout the full AI lifecycle, irrespective of where their AI applications were built or in which environment they currently run. AI OpenScale can detect and address bias across the spectrum of AI applications, as those applications are being run.

A new research project at Rochester Institute of Technology will help ensure the endangered language of the Seneca Indian Nation will be preserved. Using deep learning, a form of artificial intelligence, RIT researchers are building an automatic speech recognition application to document and transcribe the traditional language of the Seneca people. The work is also intended to be a technological resource to preserve other rare or vanishing languages.

“The motivation for this is personal. The first step in the preservation and revitalization of our language is documentation of it,” said Robert Jimerson (Seneca), a computing and information sciences doctoral student at RIT and member of the research team. He brought together tribal elders and close friends, all speakers of Seneca, to help produce audio and textual documentation of this Native American language spoken fluently by fewer than 50 individuals.

The Dark Energy Spectroscopic Instrument, scheduled to begin in 2019, will measure the velocities of about 30 million galaxies and quasars over five years. The Large Synoptic Survey Telescope, scheduled to begin in the early 2020s, will collect more than 30 terabytes of data each night—for a decade.

“The volume of datasets [from those surveys] will be at least an order of magnitude larger,” says Camille Avestruz, a postdoctoral researcher at the University of Chicago.

To keep up, astrophysicists like Schawinski and Avestruz have recruited a new class of non-scientist scientists: machines.

Researchers are using artificial intelligence to help with a variety of tasks in astronomy and cosmology, from image analysis to telescope scheduling.

Among the top challenges of reproducible computational science are: (1) creation, curation, usage and publication of research software; (2) acceptance, adoption and standardization of open-science practices; (3) misalignment with academic incentive structures and institutional processes for career progression. I will address here mainly the first two, proposing a praxis of reproducible computational science.

Built by U.K.-based startup Digital Surgery, the mobile app Touch Surgery helps medical professionals learn procedures or prepare for surgical cases with simulations and quizzes.

Now used in more than 150 U.S. residency programs, the app has a reference library of surgical maps and a virtual human patient that trains surgeons to make the right decisions at the right time during a procedure.

Digital Surgery, a member of NVIDIA’s Inception virtual accelerator program, is also developing an operating room tool called GoSurgery. It improves coordination between surgeons and their teams to manage workflows and aid in real-time operating room decisions.

datascience@berkeley created this website to show the timing of the pendulum and its speed and strength over time. Below is the split in party control of each Congress, as well as the strength of that split — by how much did each congressperson win their last election?

Amazon is gobbling up Google search results for companies even when users search specific key words related to those firms, market research firm Gartner L2 notes in a new report.

An NSF-funded pilot program led by the Ohio Supercomputer Center (OSC) seeks to grow the country’s research workforce with individuals trained in computational and data science, particularly through curriculum taught at institutions faced with resource limitations and focused on minority inclusion.

OSC staff members will coordinate the Computational and Data Science (CDS) Curriculum Exchange (C2Exchange) in collaboration with colleagues at Bethune Cookman University, Clark Atlanta University, Morgan State University, Southeastern Universities Research Association (SURA), Southern University and A&M College, and the University of Puerto Rico at Mayaguez. The three-year, $499,734 grant program was funded in June, and a project kickoff meeting was held in September.

What does a nuclear missile accident have in common with a cybersecurity or data privacy issue in a merger or integration? While it is easy to see that most people would prefer not to encounter either of them, in both situations, small missteps can lead to wide-ranging and potentially catastrophic consequences. As an M&A practitioner, however, you are much more likely to be dealing with the GDPR than an ICBM.

In fact, M&A practitioners are almost assured of encountering these issues in the course of a deal and integration. That’s because there is no area in M&A and integration practice that is as complex, rapidly changing, and risky as cybersecurity and data privacy.

… Acquiring the film in a form that allows you to pause and rewind is a necessary first step. From there, it’s all about approach. The approach I follow is one I learned when I first got my start in the basketball world. I was working for Dean Oliver, one of the leaders of the basketball statistical analysis movement, watching film and tagging each play according to a certain system. The work I was doing was monotonous, but it was very instructive. I broke down each play at an elemental level, watching to see what caused the outcome and tagging it appropriately. It trained me to think through the game, to watch each play and ask why?

I would recommend the same approach when you watch film. Watch to see what happened on a play—say, a wide-open three, for instance. Then rewind and watch again, this time trying to identify why the player was open. How did the defense move? Why did it move in that way? What were they supposed to do? One trick is to watch for player and coach reactions after plays, or listen for mics picking up their conversation. With a little guesswork, you can learn a lot about what they expected to happen.

Then think through the strategy of it all.

Writing documentation is important – it’s the first point of contact between many users and your project, and can be a pivotal moment in whether they decide to adopt your tech or become a contributor.

However, it can be a pain to iterate on documentation, as it is often involves a lot of rapid iteration locally, followed by a push to GitHub where you “just trust” that the author has done a good job of writing content, design, etc.

A really helpful tip here is to use Continuous Integration to build and preview your documentation. This allows you to generate a link to the build docs, which is a much better way of reviewing than looking at the raw text.

For the longest time, GitHub was all about storing source code and sharing it either with the rest of the world or your colleagues. Today, the company, which is in the process of being acquired by Microsoft, is taking a step in a different but related direction by launching GitHub Actions. Actions allow developers to not just host code on the platform but also run it. We’re not talking about a new cloud to rival AWS here, but instead about something more akin to a very flexible IFTTT for developers who want to automate their development workflows, whether that is sending notifications or building a full continuous integration and delivery pipeline.