Data Science newsletter – April 30, 2018

Newsletter features journalism, research papers, events, tools/software, and jobs for April 30, 2018

GROUP CURATION: N/A

 
 
Data Science News



Data Visualization of the Week

Taylor Baldwin


from


Chinese university explores links with Birmingham on big data project

Xinhua | English.news.cn


from

A delegation of senior leaders from China’s Southeast University (SEU) have visited the University of Birmingham to explore setting up a joint research institute.

Led by Executive Vice President Professor Wang Baoping, delegates from the Nanjing-based university visited Birmingham to discuss how a joint institute might benefit research collaboration in biomedical engineering and big data.

The delegates from the university met their British counterparts in computational biology and biomedics, as well as discussing collaboration opportunities in Chemical Engineering.


Senate confirms Trump’s pick for NSA, Cyber Command

POLITICO, Martin Matishak


from

The Senate Tuesday quietly confirmed President Donald Trump’s nominee to lead the National Security Agency and U.S. Cyber Command.

U.S. Army Cyber Command chief Lt. Gen. Paul Nakasone was unanimously confirmed by voice vote to serve as the “dual-hat” leader of both organizations. The two have shared a leader since the Pentagon established Cyber Command in 2009.

He will replace retiring Navy Adm. Mike Rogers after a nearly four-year term.

The Senate Intelligence and Armed Services committees both previously approved Nakasone’s nomination by voice vote.


Secrets of the $500K Amazon Alexa Prize winner: Inside the Univ. of Washington’s ‘socialbot’

GeekWire, Taylor Soper


from

How long can a robot have an intelligent conversation with a human?

That was the challenge posed to hundreds of university students last year by Amazon as part of its inaugural Alexa Prize competition, which tested the boundaries of the company’s artificial intelligence-powered voice platform, also known as Alexa.

The winning team came right out of Amazon’s backyard in Seattle, as five University of Washington students won $500,000 for its Sounding Board “socialbot” that impressed a panel of judges for its ability to hold a conversation about pop culture, news events, and more. It received an average score of 3.17 on a 5-point scale from the panel of judges and achieved an average conversation duration of 10:22.


Google’s Sergey Brin warns of the threat from AI in today’s ‘technology renaissance’

The Verge, James Vincent


from

Google co-founder Sergey Brin has warned that the current boom in artificial intelligence has created a “technology renaissance” that contains many potential threats. Writing in the company’s annual Founders’ Letter, published Friday, the Alphabet president struck a note of caution. “The new spring in artificial intelligence is the most significant development in computing in my lifetime,” writes Brin. “Every month, there are stunning new applications and transformative new techniques.” But, he adds, “such powerful tools also bring with them new questions and responsibilities.”

Brin starts his letter by quoting the opening lines of Charles Dickens’ A Tale of Two Cities — “It was the best of times, it was the worst of times.” He notes how computing power has exploded since Google was founded in 1998, and how, at that time, the technique which now forms the backbone of contemporary AI, neural networks, was just “a forgotten footnote in computer science.”


[FoR&AI] The Origins of “Artificial Intelligence”

Rodney Brooks


from

It is generally agreed that John McCarthy coined the phrase “artificial intelligence” in the written proposal2 for a 1956 Dartmouth workshop, dated August 31st, 1955. It is authored by, in listed order, John McCarthy of Dartmouth, Marvin Minsky of Harvard, Nathaniel Rochester of IBM and Claude Shannon of Bell Laboratories. Later all but Rochester would serve on the faculty at MIT, although by early in the sixties McCarthy had left to join Stanford University. The nineteen page proposal has a title page and an introductory six pages (1 through 5a), followed by individually authored sections on proposed research by the four authors. It is presumed that McCarthy wrote those first six pages which include a budget to be provided by the Rockefeller Foundation to cover 10 researchers.


The promise and peril of military applications of artificial intelligence

Bulletin of the Atomic Scientists, Michael Horowitz


from

The potential promise of AI—including its ability to improve the speed and accuracy of everything from logistics to battlefield planning and to help improve human decision-making—is driving militaries around the world to accelerate their research into and development of AI applications. For the US military, AI offers a new avenue to sustain its military superiority while potentially reducing costs and risk to US soldiers. For others, especially Russia and China, AI offers something potentially even more valuable—the ability to disrupt US military superiority. National competition in AI leadership is as much or more an issue of economic competition and leadership than anything else, but the potential military impact is also clear. There is significant uncertainty about the pace and trajectory of artificial intelligence research, which means it is always possible that the promise of AI will turn into more hype than reality. Moreover, safety and reliability concerns could limit the ways that militaries choose to employ AI.


Taskonomy – Disentangling Task Transfer Learning

Amir R. Zamir, Alexander Sax, William B. Shen, Leonidas Guibas, Jitendra Malik, Silvio Savarese


from

Do visual tasks have a relationship, or are they unrelated? For instance, could having surface normals simplify estimating the depth of an image? Intuition answers these questions positively, implying existence of a “structure” among visual tasks. Knowing this structure has notable values; it is the concept underlying transfer learning and provides a principled way for identifying redundancies across tasks, e.g., to seamlessly reuse supervision among related tasks or solve many tasks in one system without piling up the complexity.

We propose a fully computational approach for modeling the structure of the space of visual tasks.


Uru Team Joins Adobe

Uru, Adobe


from

At Uru, we embarked on our AI and computer vision journey two years ago with the goal of helping brands understand and leverage all the visual content being created today. Now, the AI-powered revolution that once seemed like science fiction is transforming our personal and professional lives – including how brands interact with us in real-time.

We’ve always been looking around the corner at what’s next with AI and innovation, and who can help us stay ahead of the curve. Today, the Uru team is excited to announce that it is joining Adobe as part of the group working on Adobe Sensei – the company’s AI and machine learning platform.


Weekly Wrapup: Insurers reach the next phase of digital transformation

Digital Insurance, Nathan Golia


from

The Global Insurance Symposium isn’t a technology conference per se, but it’s getting harder than ever for conversations in the insurance industry to focus on anything else.

The three-day Des Moines event featured speakers from around the world and across the insurance industry — from business-side executives to state regulators to consultants to representatives from the insurtech sector — dishing on the transformative impact of digital on the sector. But these weren’t 30,000-foot views of what might come. Carriers are discussing in-market initiatives that paint a very different picture of insurance than the traditional staid one.

In giving an update about the life insurer’s wearable-technology powered wellness initiative, Vitality, Matthew Gabriel, AVP of innovation at John Hancock, said that his company has learned to deliver real value to customers in order to get them to adopt the program. An audience question pressed him to compare the adoption of wearables to telematics in auto insurance, which was slow to grow. Gabriel credited Vitality’s successes to having a conversation with customers 20 to 30 times per month about topics that don’t directly relate to their premium rate or bill.


Artificial intelligence helps Soldiers learn many times faster in combat

U.S. Army Research Lab


from

New technology allows U.S. Soldiers to learn 13 times faster than conventional methods and Army researchers said this may help save lives.

At the U.S. Army Research Laboratory, scientists are improving the rate of learning even with limited resources. It’s possible to help Soldiers decipher hints of information faster and more quickly deploy solutions, such as recognizing threats like a vehicle-borne improvised explosive device, or potential danger zones from aerial war zone images.

The researchers relied on low-cost, lightweight hardware and implemented collaborative filtering, a well-known machine learning technique on a state-of-the-art, low-power Field Programmable Gate Array platform to achieve a 13.3 times speedup of training compared to a state-of-the-art optimized multi-core system and 12.7 times speedup for optimized GPU systems.


Foxconn Will Drain 7 Million Gallons of Water Per Day From Lake Michigan to Make LCD Screens

Gizmodo, AJ Dellinger


from

This week, the Wisconsin Department of Natural Resources gave the go-ahead to Taiwanese tech manufacturer Foxconn to siphon off seven million gallons of water per day from Lake Michigan, despite protests from conservation groups.

The massive diversion of water from the lake will be used to produce LCD screens at the company’s planned $10 billion, 20 million square foot manufacturing plant set to be built in Mount Pleasant, Wisconsin.

Nearly 2.7 million gallons of the water—about 39 percent of the daily intake from the factory—will be lost in the process, primarily from evaporation. The remaining water will be treated and returned to the lake basin.


Satellites, supercomputers provide real-time crop data

Quincy (IL) Herald-Whig, Deborah Gertz Husar


from

Corn and soybean fields look similar from space — at least they used to — but scientists have proven a new technique for distinguishing the two crops using satellite data and the processing power of supercomputers.

“If we want to predict corn or soybean production for Illinois or the entire United States, we have to know where they are being grown,” said Kaiyu Guan, assistant professor in the Department of Natural Resources and Environmental Sciences at the University of Illinois, Blue Waters professor at the National Center for Supercomputing Applications and the new study’s principal investigator.

The advancement, published in Remote Sensing of Environment, is a breakthrough because national corn and soybean acreages previously were only made available to the public four to six months after harvest by the USDA. The lag meant policy decisions were based on stale data. But the new technique can distinguish the two major crops with 95 percent accuracy by the end of July for each field — just two or three months after planting and well before harvest.


The Best Ways to Fix College Admissions Are Probably Illegal

The Atlantic, Jeffrey Selingo


from

ear after year, the admissions process at selective colleges seems to make high-schoolers and their parents only more anxious. The numbers are wild: Harvard admitted just 4.6 percent of its nearly 43,000 applicants for the class that begins this fall. Stanford accepted only 4.29 percent, and Princeton 5.5 percent. Although selective schools—those that accept fewer than half of applicants—enroll only about one-fifth of U.S. undergraduates, they account for more than one-third of applications each year.

Plenty of ideas to fix the system—to make it more bearable for students, parents, and even colleges themselves—have been floated in recent years, including restructuring the whole process to be a somewhat randomized lottery, or implementing a matching system akin to how medical-school graduates are placed in residencies. They are promising, but they have something problematic in common: In all likelihood, they’d be illegal.


Investigators searched a million people’s DNA to find Golden State serial killer

MIT Technology Review, Antonio Regalado


from

Investigators may have compared a serial killer’s DNA with that of one million unwitting genealogy enthusiasts as part of an investigation that led to the arrest earlier this week of a man accused of being California’s elusive “Golden State Killer.”

“I had no knowledge this was happening,” says Curtis Rogers, co-creator of GEDMatch, an ancestry site that a police source yesterday identified as one of those employed by investigators.

Officials in California said they had found and arrested Joseph James DeAngelo, 72, for a fearsome series of murders and rapes between 1974 and 1986 after using commercial genealogy websites, including GEDMatch, to locate one of his relatives.

GEDMatch, a no-frills website that has never advertised, is used by amateur and professional genealogists to upload and compare DNA tests, effectively crowdsourcing vast family trees.

 
Events



Voice 2018

Modev


from

Newark, NJ July 24-26. “Natural language is revolutionizing the way we interact with devices, services, products and one another. VOICE is where the world’s top platform providers, brands, agencies, investors, startups and developers explore the cutting edge of the multi-modal engagement era.” [$$$]


The Challenges and Opportunities of Explainable AI

Intel


from

San Francisco, CA May 22, starting at 6 p.m., Galvanize (44 Tehama St.) “As part of Intel AI’s technical panel series for entrepreneurial leaders to share their views on important emerging topics, Casimir Wierzynski (Senior Director of Research) invites you to an upcoming discussion he is moderating on ‘The Challenges and Opportunities of Explainable AI.'” [registration required]

 
Deadlines



2018 BIDS Data Science Faire – Call for Posters/Demos

Berkeley, CA May 8, starting at 1:30 p.m., Berkeley Institute for Data Science (190 Doe Library). Deadline for submissions is May 1.

Call for Papers :: FAT ML

Stockholm, Sweden 5th Workshop on Fairness, Accountability, and Transparency in Machine Learning, Co-located with ICML 2018 on July 14-15. Deadline for submissions is May 1.
 
Moore-Sloan Data Science Environment News



Dr. Julia Kempe appointed as Director of the NYU Center for Data Science

Medium, NYU Center for Data Science


from

Dr. Kempe comes with two decades of in-depth experience in both academia and data-driven industry. She is currently a Senior Researcher at a leading hedge fund that trades in markets around the world, employing complex mathematical models and tools from machine learning and statistics to analyze vast amounts of data.

Prior to working in finance, she held academic appointments as Junior and then Senior Researcher at the CNRS in France since 2001, and as Professor of Computer Science at Tel-Aviv University (2007–2011).

 
Tools & Resources



Value-Suppressing Uncertainty Palettes (VSUPs)

Michael Correll, Dominik Moritz, Jeffrey Heer


from

Understanding uncertainty is critical for many analytical tasks. One common approach is to encode data values and uncertainty values independently, using two visual variables. These resulting bivariate maps can be difficult to interpret, and interference between visual channels can reduce the discriminability of marks. To address this issue, we contribute Value-Suppressing Uncertainty Palettes (VSUPs). VSUPs allocate larger ranges of a visual channel to data when uncertainty is low, and smaller ranges when uncertainty is high. This non-uniform budgeting of the visual channels makes more economical use of the limited visual encoding space when uncertainty is low, and encourages more cautious decision-making when uncertainty is high.


NVIDIA SMP Assist API for VR Programming

NVIDIA Developer Blog, Vishwesh Inamdar |


from

“The NVIDIA SMP (simultaneous multi-projection) Assist NVAPI driver extension is a simple method for integrating Multi-Res Shading and Lens-Matched Shading into a VR application. It encapsulates a notable amount of state setup and API calls into a simplified API, thereby substantially reducing the complexity of integrating NVIDIA VRWorks into an application. Specifically, the SMP Assist driver handles creating and managing fast geometry shaders, viewport state, and scissor state in lieu of managing these states manually in the application.”


Tips for High Availability

Medium, Netflix Technology Blog, Andy Glover and Katharina Probst


from

“At Netflix, we have built and use Spinnaker as a platform for continuous integration and delivery. Many of the best practices discussed here have been encoded into Spinnaker, so that they are easy to follow. While in this article we show how we internally encode the best practices in Spinnaker, the tips and best practices are more general and will help anyone make their systems be highly available.”

 
Careers


Postdocs

Post-Doctoral Fellowship



University of Washington, Friday Harbor Laboratories; San Juan Island, WA

Postdoctoral Scholar in Computational Neuroscience



Penn State University; State College, PA
Full-time positions outside academia

UX Researcher #0409



PARC; Palo Alto, CA
Internships and other temporary positions

Deep Learning Intern



Dolby; San Francisco, CA

Leave a Comment

Your email address will not be published.