Data Science newsletter – March 19, 2021

Numerous proposals, prototypes, and models have emerged for using artificial intelligence (AI) and machine learning to predict individual risk related to covid-19. In the United States, for instance, the Department of Veterans Affairs uses individualised risk scores to allocate medical resources to people with covid-19,1 and prisons have sought to detect symptoms by processing inmates’ phone calls.2 Further tools, such as vulnerability predictions for individuals3 and voice based detection of infection,4 are on the horizon. But use of AI for such purposes has given rise to questions about legality.

When a state or federal government seeks to use AI models to predict an individual’s risk of covid-19, the key legal questions will ultimately turn on how effective the models are and how much they burden legal interests. We focus on two of the most salient legal concerns under US law: privacy and discrimination. Challenges on privacy or discrimination grounds might appear in a variety of contexts, including challenges to regulatory decisions, tort actions, or lawsuits under health privacy laws. We argue that the basic need to balance benefits against burdens runs through all of these legal regimes. Governments implementing risk scoring tools must show that their tools produce valid, reliable predictions and burden individuals’ civil liberties no more than necessary. In evaluating the legality of public health use of algorithms, courts will likely also probe how the output of these tools is used to shape policies and programs. But showing that a model performs well and does not exceedingly burden privacy and other interests are essential preconditions for lawful deployment.

Google has introduced its latest Nest Hub, with a focus on making it the “ultimate bedside assistant and alarm clock”.

Speaking with media about the “second-generation” Nest Hub, Nest product manager Ashton Udall said it came as a surprise to Google that people were mostly using the Nest Hub in their bedroom, so the company decided to double down and focus on adding new bedroom-specific features.

“When we talked to people, what they really wanted more help with in their bedrooms, by far and away, the resounding response was ‘help me get better sleep’,” he said.

“Quality of sleep has become the number one concern when it comes to health and wellness in the home … one in three adults are sleep deprived, one in seven have a sleep breathing disorder … and one in two adults, myself included, have trouble falling asleep.”

The auction rules were simple: After each bid, the four companies had an hour to raise the buying price by at least a million dollars. This hour-​long countdown started at the time stamped on the email holding the latest bid, and at the end of the hour, if no one lodged a new bid, the auction was over. DeepMind bid with company shares, not cash, but it couldn’t compete with the giants and soon dropped out. That left Baidu, Google, and Microsoft. As the bids continued to climb, first to $15 million and then to $20 million, Microsoft dropped out too, but then returned. Each increment felt heavy with meaning as [Geoff] Hinton and his students debated which company they’d rather join. Late one afternoon, as they looked out the window, two airplanes flew past from opposite directions, leaving contrails that crossed in the sky like a giant X right above a set of mountain peaks. Punchy with excitement, they mused about what this might mean, before remembering that Google was headquartered in a place called Mountain View. “Does that mean we should join Google?” Hinton asked. “Or does it mean we shouldn’t?”

At about $22 million, Hinton temporarily suspended the auction to hold a discussion with one of the bidders, and half an hour later Microsoft dropped out again. That left Baidu and Google, and as the hours passed, the two companies took the price still higher. Kai Yu handled the initial Baidu bids, but when the price reached $24 million, a Baidu executive took over from Beijing.

The climate analytics firms often shield their data models from public scrutiny as proprietary information, unlike the computer models that academic and government researchers typically rely on for their less granular projections of the warming planet. That makes it impossible to independently validate their work, scientists say.

“Do these guys know what they are doing? I’m not convinced that they do,” said Upmanu Lall, director of the Columbia Water Center at Columbia University, who has reviewed some firms’ methodologies. “Your models are garbage. And, unfortunately, that’s a problem.”

“It’s not that [they] have some special sauce,” said Rutgers University climate scientist Robert Kopp, who contributes to climate analytics service firms Rhodium Group and First Street Foundation, which publish their methodologies. “[They] don’t want to talk about what you’re doing.”

Digit marches on two legs across the floor of the University of Michigan’s Ford Motor Co. Robotics Building, while Mini-Cheetah — staccato-like — does the same on four and the yellow-legged Cassie steps deliberately side-to-side.

A grand opening was held Tuesday for the four-story, $75 million, 134,000-square-foot (11,429-square-meter) complex. Three floors house classrooms and research labs for robots that fly, walk, roll and augment the human body.

On the top floor are Ford researchers and engineers and the automaker’s first robotics and mobility research lab on a university campus.

Two years after a high-profile falling out, the University of California (UC) system and the academic publishing giant Elsevier have patched up differences and agreed on what will be the largest deal for open-access publishing in scholarly journals in North America. The deal is also the world’s first such contract that includes Elsevier’s highly selective flagship journals Cell and The Lancet.

The deal meets demands made by UC when it suspended negotiations with Elsevier in 2019. It allows UC faculty and students to read articles in almost all of Elsevier’s more than 2600 journals, and it enables UC authors to publish articles that they can make open access, or free for anyone to read, by paying a per-article fee. Elsevier says it will discount those open-access fees, and UC says it will subsidize their authors.

UC estimates the new deal will cost its libraries’ budget 7% less than what they would have paid had it extended its old contract with Elsevier.

… Only a fourth of people with hypertension have it under control, in part because sphygmomanometers, whether used in a doctor’s office or via clunky home units, don’t supply a steady stream of readings, multiple times a day and in different settings, to help determine the proper treatment. A new company, launching today, thinks it can improve those grim statistics. It’s called Riva Health, chosen as a tribute to the Italian inventor. Riva measures blood pressure from the arteries of one’s fingertip, which it captures using the flash of an iPhone’s camera.

UMass Amherst researchers found that the computing power needed to train a large AI model can produce over 600,000 pounds of CO2 emissions – that’s five times the amount of the typical car over its lifespan! These models often take even more energy to process in real-world production settings (otherwise known as the inference phase). NVIDIA estimates that 80-90 percent of the cost incurred from running a neural network model comes during inference, rather than training.

To make more progress in the AI field, popular opinion suggests we’ll have to make a huge environmental tradeoff. But that’s not the case. Big models can be shrunk down to size to run on an everyday workstation or server, without having to sacrifice accuracy and speed. But first, let’s look at why machine learning models got so big in the first place.

PatSnap, which was founded in 2007 by Jeffrey Tiong, began as a directory for intellectual property (IP), helping enterprises pull in data for R&D and ideation purposes. Since then, it’s developed technologies in the AI subfields of machine learning, natural language processing, and computer vision that analyze and identify the relationships between millions of unstructured data points across disparate IP sources. PatSnap claims this can help deliver insights that guide R&D decisions and help to shorten the time it takes to bring new creations to market.

Shortly before Ibram X. Kendi arrived from Washington, D.C., last June to launch the Boston University Center for Antiracist Research, he was introduced by University President Robert A. Brown to the Boston Globe’s new editorial page editor, Bina Venkataraman, via email. The antiracist scholar and the prominent opinion journalist hit it off—discovering a shared interest in Boston’s history as a hub of 19th-century anti-slavery publications, notably William Lloyd Garrison’s abolitionist newspaper The Liberator.

On Tuesday, Kendi and Venkataraman announced bold plans to reimagine those 19th-century printed journals for a 21st-century digital audience. The center is collaborating with the Globe editorial team, Globe Opinion, to publish The Emancipator, an independent antiracist multimedia platform. Antiracism scholar Kendi, who holds BU’s Andrew W. Mellon Professorship in the Humanities and is also a College of Arts & Sciences professor of history, and Venkataraman, who served as a science policy expert in the Obama White House, are the cofounders of the new platform. The mission, they say, is to help reframe today’s national conversation on race.

Amazon and the Haslam family are making a significant donation to the University of Tennessee at Knoxville’s Haslam College of Business to establish a new professor position.

Amazon and the Haslam family each donated $750,000, for a total of $1.5 million, to create an endowed distinguished professorship in business analytics and data science, according to a press release from Amazon.

After spending a year building one of the only U.S. pandemic-data sources, we have come to see the government’s initial failure here as the fault on which the entire catastrophe pivots. The government has made progress since May; it is finally able to track pandemic data. Yet some underlying failures remain unfixed. The same calamity could happen again.

Data might seem like an overly technical obsession, an oddly nerdy scapegoat on which to hang the deaths of half a million Americans. But data are how our leaders apprehend reality. In a sense, data are the federal government’s reality. As a gap opened between the data that leaders imagined should exist and the data that actually did exist, it swallowed the country’s pandemic planning and response.

The European Commission will unveil its AI rules this spring, requiring “high-risk” AI systems to meet minimum standards regarding trustworthiness. But with European countries already struggling to address racial discrimination when it comes to representation, policing and online abuse, Sarah Chander of digital rights group EDRi said those issues are likely to seep into tech.

“We shouldn’t see the issues of the potential harmful impact on racialized communities through tech as a U.S. issue. It’s going to be wherever you find manifest structural discrimination and racial inequality,” Chander said, adding that complacent attitudes among policymakers and an unwillingness to recognize the problem will only exacerbate it.

The Association for Computational Linguistics has created ACLRollingReview, a system that is intended to be used in the long term by all *ACL conferences for reviewing. We have appointed three inaugural editors in chief of ACLRollingReview:

The head of the CDC warned of a COVID-19 surge over spring break.

The US had more air travelers Friday than on any other day in the past year.

“All in the context of still 50,000 cases per day,” Dr. Rochelle Walensky said.

“Here I am, writing the first edition of my ‘Weekly Awesome Tricks And Best Practices From Kaggle’. Throughout the series, you will find me writing about anything that can be useful during a typical data science workflow including code shortcuts related to common libraries, best practices that are followed by top industry experts on Kaggle, and so on, all learned by me during the past week. Enjoy!”

Cosmos is a computing platform that combines the best aspects of microservices with asynchronous workflows and serverless functions. Its sweet spot is applications that involve resource-intensive algorithms coordinated via complex, hierarchical workflows that last anywhere from minutes to years. It supports both high throughput services that consume hundreds of thousands of CPUs at a time, and latency-sensitive workloads where humans are waiting for the results of a computation.

This article will explain why we built Cosmos, how it works and share some of the things we have learned along the way.