Data Science newsletter – August 8, 2019

GROUP CURATION: N/A

Most of us are similar with ‘viruses’ and malware relating to our computers or smartphones, but Yoshihiro Minagawa, a researcher from the University of Tokyo has taken it on literally – he has invented a portable, low-cost, battery-powered device that pairs with a smartphone, which was tested with viruses but could also detect other biological markers. His initial findings, together with other teammates were published recently in the journal, Lab on a Chip.

The current leading method to assess the presence of viruses and other biological markers of disease is effective but large and expensive. It is prohibitively difficult for use in many situations, especially due to certain economic and geographic factors. Although highly accurate at counting viruses, these tools are just too cumbersome for many situations, especially when rapid diagnosis is required.

Foreign-born Ph.D. graduates with science and engineering degrees from American universities apply to and receive offers for technology startup jobs at the same rate as U.S. citizens, but are only half as likely to actually work at fledgling companies, a Cornell study has found.

Instead, they choose to work at large technology companies with the resources and experience to sponsor foreign workers for highly coveted H-1B or permanent residency visas.

The findings suggest visa policies are an obstacle for small companies seeking to hire foreign-born workers with specialized, in-demand skills. Because so many Ph.D. graduates in STEM fields are international – more than half in some disciplines – this creates an uneven playing field for startups competing with established companies to attract top talent.

University of Maryland researchers created 1,213 questions in collaboration with computers to identify flaws in machine-learning language models

The holy grail of artificial intelligence is a machine that truly understands human language and interprets meaning from complex, nuanced passages. When IBM’s Watson computer beat famed “Jeopardy!” champion Ken Jennings in 2011, it seemed as if that milestone had been met. However, anyone who has tried to have a conversation with virtual assistant Siri knows that computers have a long way to go to truly understand human language. To get better at understanding language, computer systems must train using questions that challenge them and reflect the full complexity of human language.

Researchers from the University of Maryland have figured out how to reliably create such questions through a human-computer collaboration, developing a dataset of more than 1,200 questions that, while easy for people to answer, stump the best computer answering systems today. The system that learns to master these questions will have a better understanding of language than any system currently in existence. The work is described in an article published in the journal Transactions of the Association for Computational Linguistics.

Celect’s cloud analytics platform enables retailers to optimize inventory across channels. Nike plans to use that data science knowhow for its Consumer Direct Offense strategy.

The 25th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) opened on August 4 in Anchorage, Alaska. KDD 2019 features both a Research track and an Applied Data Science (ADS) track, and today organizers announced their Best Paper Awards. A Cornell University research team took top honours in the research track for Network Density of States; while in the ADS track the winner was Actions Speak Louder than Goals: Valuing Player Actions in Soccer, from researchers at Katholieke Universiteit Leuven and SciSports.

“This is the worlds foremost data scientist conference,” he said. “Where the researchers and the practitioners come together to find out the latest advances and how they can be put to use.” … The Municipality of Anchorage teamed up with UAA, Visit Anchorage and Axiom Data to apply as a location for the event in 2016 according to Brendan Babb, the municipality’s Chief Innovation Officer. Babb says the attendance at this year’s conference should translate to roughly $5 million dollars in direct economic impacts, as well as an estimated $300,000 in municipal taxes through hotel and rental car taxes. He also hopes the successful bid will lead to increased attention for data collections that are unique to Alaska.

Which hardware platforms — TPU, GPU or CPU — are best suited for training deep learning models has been a matter of discussion in the AI community for years. A new Harvard University study proposes a benchmark suite to analyze the pros and cons of each.

“ParaDnn” is the product of a group of Harvard University researchers at the School of Engineering and Applied Sciences. The parameterized benchmark suite conducts subjective comparisons between different computing platforms for deep learning models.

Unlike recent benchmarking efforts that have been limited to relatively small collections of DNN models, ParaDnn was designed to support broad and comprehensive benchmark suites. It can seamlessly generate thousands of multi-layer models, including fully-connected models (FD), convolutional neural networks (CNN) and recurrent neural networks (RNN). Below are the ranges of the hyperparameters and dataset variables for each of the models.

Milind Tambe, a leader in applying Artificial Intelligence to address pressing societal challenges, has joined the faculty at the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) as the Gordon McKay Professor of Computer Science and Director of the Center for Research on Computation and Society (CRCS).

Tambe’s research explores fundamental problems in computational game theory, machine learning, automated planning, intelligent agents, and multi-agent interactions. He comes to SEAS from the University of Southern California, where he was a professor in engineering and the Founding Co-Director of the USC Center for Artificial Intelligence in Society.

When [Keenan] Robinson arrived at Georgia State University in Atlanta, he wanted to major in nursing. “I always knew I had a passion for helping people,” he said. Biology had been his best subject in high school. “My dad, my mom would always kind of call me like the king of trivia because I’d always have just like random science facts,” he said.

During his freshman year, Robinson earned a B average. But the university was closely tracking his academic performance and knew from 10 years of student records that Robinson wasn’t likely to make the cut for the nursing program.

Georgia State is one of a growing number of schools that have turned to big data to help them identify students who might be struggling – or soon be struggling – academically so the school can provide support before students drop out.

Artificial intelligence experts have long warned that algorithmic decision-making systems can be used to discriminate. Now, a leaked memo reveals a new rule proposed by the Department of Housing and Urban Development (HUD) would make it much easier for banks, insurance companies, and landlords to use discriminatory algorithms without the threat of lawsuits.

The new rule takes aim at a 2015 Supreme Court ruling, which decided that consumers could combat housing discriminatory business practices by making “disparate-impact claims” under the Fair Housing Act of 1968. In a disparate-impact claim, if you find out that a business practice had a disproportionate effect on certain groups of people, then you can hold that business liable—even if it was an unintended consequence. Claimants, then, could not only challenge discriminatory housing practices but use statistical analysis of past behavior to prove that the discrimination was happening and detrimental to certain groups.

The White House launched the BRAIN Initiative on April 2, 2013, with the goal of supporting the development and application of innovative technologies that can create a dynamic understanding of brain function. The initiative has supported more than $1 billion in research and has led to new insights, new drugs, and new technologies to help individuals with brain disorders.

But this wealth of research comes with challenges, according to Russell Poldrack, a psychology professor with a computing bent at Stanford University. Psychology and neuroscience struggle to build on the knowledge of its disparate researchers.

“Science is meant to be cumulative, but both methodological and conceptual problems have impeded cumulative progress in psychological science,” Poldrack and collaborators from Stanford, Dartmouth College and Arizona State University wrote in a Nature Communications paper out in May 2019.

Because of its large number of operational CCTV cameras, London is now at the forefront of a battle over the use of facial recognition by the authorities that’s escalating across many democratic countries. The Romford operation was one of 10 such events around London carried out by the Met police over a period of three years, including twice at Notting Hill Carnival. Over the course of the trials, police planned to gather evidence about the accuracy and bias of the system and to assess whether the use of the dragnet technology could be justified by its potential benefits — preventing or solving major acts of violent crime.

The ride-hailing giants released a joint analysis showing that their vehicles are responsible for significant portions of VMT in six major urban centers. Still, Uber and Lyft’s combined share is still vastly outstripped by personal vehicles. As Chris Pangilinan, Uber’s head of global policy for public transportation, wrote in a blog post accompanying the findings, “although TNCs are likely contributing to an increase in congestion, its scale is dwarfed by that of private cars and commercial traffic.”

Led by the respected transportation consultancy Fehr & Peers, the analysis provides a high-level view of the combined mileage contributions from Uber and Lyft, as a share of overall VMT, over a recent month in the Boston, Chicago, L.A., San Francisco, Seattle, and Washington, D.C. areas. Results are shown at the level of the larger metropolitan landscape, which includes both the central city and its surrounding suburbs, as well as the level of the core county that contains the city’s most concentrated homes and jobs.

A study looking at the cost of living in 257 cities across the United States has ranked Seattle as the fifth most expensive place to live in the country.

Seattle trails only New York City (Manhattan), San Francisco, Honolulu and Brooklyn in the latest Cost of Living Index from The Council for Community and Economic Research. Portland, Ore., made the top 25 along with cities such as Boston, Washington, D.C., San Diego and Chicago.

When police partner with Ring, Amazon’s home surveillance camera company, they get access to the “Law Enforcement Neighborhood Portal,” an interactive map that allows officers to request footage directly from camera owners. Police don’t need a warrant to request this footage, but they do need permission from camera owners.

Emails and documents obtained by Motherboard reveal that people aren’t always willing to provide police with their Ring camera footage. However, Ring works with law enforcement and gives them advice on how to persuade people to give them footage.

Cleveland, OH October 18-20 at NASA Glenn Research Center. [registration required]

New York, NY November 12-13. “NEXT NYC drives forward-thinking conversations, highlights influential voices, and ultimately, spurs real action to improve patient lives.” [registration required]

Who knew Minecraft offered such a rich training ground for AI and machine learning algorithms? Earlier this month, Facebook researchers posited that the hit game’s constraints make it well-suited to natural language understanding experiments. And in a newly published paper, a team at Carnegie Mellon describe a 130GB-734GB corpus intended to inform AI development — MineRL — that contains over 60 million annotated state-action pairs (recorded over 500 hours) across a variety of related Minecraft tasks, alongside a novel data collection scheme that allows for the addition of tasks and the gathering of complete state information suitable for “a variety of methods.”

If you have responses to customer surveys, social media comments, or product reviews in Google Sheets, you can automatically analyze this data in that very same spreadsheet. Not only is it easy to get started with text analysis for Google Sheets, but it will also save you precious time and resources.

In this post, I compare some of the popular models of integrating data science teams within organizations. In determining the best model, I try to minimize the following unfavorable outcomes.