Data Science newsletter – October 15, 2019

GROUP CURATION: N/A

A federal appeals court on Monday rejected LinkedIn’s effort to stop a San Francisco company from using information that users of the professional networking website have deemed public.

Medicine X has been clearly held its own in the growing debate Stateside and globally about the partnership with patients to do better research and care. The degree to which its philosophy and approach has yet to pervade the mechanics of state and federal research and care systems and organizations is not always clear, but there are signs that the influence is increasing and has massive potential.

We found it bolder and more convincing than PCORI, which in spite of its vast amounts of money seems constrained by its own rules and regulations and its lack of courage in ultimately delegating control to patients. PCORI is bound very much by the conventions of scientific rigour. Medicine X seems intent on maintaining scientific validity while challenging the outdated traditions that prevent patients from being full partners in research and healthcare.

Machines’ ability to learn by processing data gleaned from sensors underlies automated vehicles, medical devices and a host of other emerging technologies. But that learning ability leaves systems vulnerable to hackers in unexpected ways, researchers at Princeton University have found.

Social media, cameras, sensors and more generate huge amounts of data that can overwhelm analysts sifting through it all for meaningful, actionable information to provide decision-makers such as political leaders and field commanders responding to security threats.

Sandia National Laboratories researchers are working to lessen that burden by developing the science to gather insights from data in nearly real time.

“The amount of data produced by sensors and social media is booming — every day there’s about 2.5 quintillion (or 2.5 billion billion) bytes of data generated,” said Tian Ma, a Sandia computer scientist and project co-lead. “About 90% of all data has been generated in the last two years — there’s more data than we have people to analyze. Intelligence communities are basically overwhelmed, and the problem is that you end up with a lot of data sitting on disks that could get overlooked.”

Google Stadia will be faster and more responsive than local gaming systems in “a year or two,” according to VP of engineering Madj Bakar. Thanks to some precog trickery, Google believes its streaming system will be faster than the gaming systems of the near-future, no matter how powerful they may become. But if the system is playing itself, does that really count?

The Faculty Council has begun considering scheduling options for the new School of Engineering and Applied Sciences campus in Allston, due to open next fall. Unfortunately, past significant logistical changes, such as last year’s implementation of the new academic course schedule, cannot be said to have gone especially smoothly. Given that, we hope the administration has taken special care to learn from and avoid the mistakes of that process as it turns to scheduling for the Allston campus.

Integrating Allston into the College next year will require a number of complex logistical maneuvers to function well if it is not to worsen the learning experience of students attending classes there. As many students’ aversion over the distance to the Quad demonstrates, geography is not a problem to underestimate or under-plan for. And the Allston campus raises other problems, like guaranteeing students timely access to food and minimizing the number of trips necessary in a given day.

But ultimately, our biggest worry is that the College and the Faculty of Arts and Sciences is spreading itself too thin.

Visa, Mastercard, eBay, Stripe, and Mercado Pago have all withdrawn from the Libra Association, dealing a major blow to Facebook’s plans for a distributed, global cryptocurrency. The withdrawals were first reported by the Financial Times and Bloomberg.

A Visa spokesperson told The Verge. “Visa has decided not to join the Libra Association at this time,” the spokesperson said. “We will continue to evaluate and our ultimate decision will be determined by a number of factors, including the Association’s ability to fully satisfy all requisite regulatory expectations.”

Pinterest says it’s using machine learning techniques to identify and hide content that displays, rationalizes, or encourages self-injury. (Pinterest, of course, has a robust AI toolset at its disposal — it recently revealed that Lens, its online/offline visual search tool that taps AI to identify things captured from Pins or by a smartphone and suggest related themes and products, can now recognize 2.5 billion home and fashion objects.) The company says it has achieved an 88% reduction in reports of self-harm content by users and that it’s now able to remove such content 3 times faster.

Engineers at the University of California San Diego have developed a way to build soft robots that are compact, portable and multifunctional. The advance was made possible by creating soft, tubular actuators whose movements are electrically controlled, which makes them easy to integrate with small electronic components.

As a proof of concept, engineers used these new actuators to build a soft, battery-powered robot that can walk untethered on flat surfaces and move objects. They also built a soft gripper that can grasp and pick up small objects.

For more than a decade, architect and educator Jenny Sabin has created a series of large sculptural installations at the intersection of architecture, art and science. Integrating new technologies, inventive material and a combination of biological phenomena and mathematics, her latest work adds artificial intelligence, or AI.

“Ada,” a responsive, photoluminescent fiber pavilion designed to “smile back at you,” has just opened, suspended in a light-filled atrium at Microsoft Research in Redmond, Washington.

New reports suggest that drone activity at the southern border is spreading to nearby cities, erasing the line between police procedures and immigration enforcement.

In the first week of classes, students taking Computer Science 201 walked into the Bryan Center’s Griffith Film Theater, which has a seating capacity of 500, to find the lecture nearly packed from the first row to the last. Laptops were wide open, creating a sea of screens that lit up the theater as students waited for Owen Astrachan, professor of the practice of computer science, to begin the first lecture.

The packed class was not an exception for computer science, which is the most popular major at Duke since at least Fall 2018. There are 734 students majoring in computer science currently, wrote Camelia Pierson Eaves, undergraduate program coordinator for the department, in an email. She added that 1,746 undergraduate and 376 graduate students are enrolled in computer science courses (which likely includes students enrolled in more than one course in the department.)

Karen Willcox, University of Texas, Austin; SFI In this talk I will discuss challenges and needs for tackling some of society’s most pressing problems across science, engineering and medicine. I’ll make the case for “predictive data science” — recognizing that in high-consequence decisions, methods based on data alone are not enough. Rather we need a synergistic combination of modern data-driven and more classical physics-based perspectives. In our excitement to embrace machine learning and artificial intelligence, it is critical that we not forget about the predictive power, the interpretability, and the domain knowledge associated with physics-based models. I will also discuss implications and needs for interdisciplinary education. [video, 48:41]

Grammarly, the Kyiv, San Francisco, Vancouver, and New York startup behind the eponymous cross-platform grammar- and spell-checking platform, today revealed that it’s secured $90 million in its second-ever funding round, led by General Catalyst, with participation from IVP as well as undisclosed existing and new investors. CEO Brad Hoover said the fresh capital, which brings Grammarly’s total raised to date to roughly $200 million following a $110 million raise in May 2017, will be used to advance its technology to build a full-fledged communications assistant. He also said it’ll enable Grammarly to expand the size of its workforce globally, which now numbers north of 200 employees.

We are pleased to report that the campus has re-connected to PG&E power as of 1 p.m. today, welcome news that has brought the campus back to normal power.

Los Angeles, CA October 17-24 at Broad Art Center. “The eleven works featured in CONTEXT COLLAPSE undertake a conversation around the shifting nature of context in contemporary life and art.” [free]

“To the best of my knowledge, this is the first public implementation of Xie et al.’s 2014 OptimalPLR and GreedyPLR algorithms.”

“The U.S. Census Bureau has added 205 unique sources of data from around the world to its HIV/AIDS Surveillance Data Base, providing easy access to key findings in one place.”

PyTorch 1.3 “includes experimental support for features such as seamless model deployment to mobile devices, model quantization for better performance at inference time, and front-end improvements, like the ability to name tensors and create clearer code with less need for inline comments.”

Multimodal learning consolidates a series of disconnected, heterogenous data from various sensors and data inputs into a single model. Unlike traditional unimodal learning systems, multimodal systems can carry complementary information about each other, which will only become evident when they are both included in the learning process. Therefore, learning-based methods that combine signals from different modalities are capable of generating more robust inference, or even new insights, which would be impossible in a unimodal system.