Data Science newsletter – March 9, 2021

The Vanderbilt Data Science Institute accelerates data-driven research, promotes collaboration and trains future leaders. The institute brings together experts in data science methodologies with leaders in all academic disciplines to spark discoveries and to study the impact of big data on society. The institute is educating students in computational and statistical data science techniques to become future leaders in industry, government, academia and the nonprofit sector. [sizzle reel, 3:23]

In November, Springer Nature, one of the world’s largest publishers of scientific journals, made an attention-grabbing announcement: More than 30 of its most prestigious journals, including the flagship Nature, will now allow authors to pay a fee of $11,390 to make their papers freely available for anyone to read online.

This move, by a company that publishes more than 3,000 academic titles, has been hailed as a landmark step — and a victory for an open-access movement that seeks to supplant the traditional subscription-based model of academic publishing. And at first glance, Springer Nature’s open-access option appears to be a positive development. Most scientific articles are paywalled, accessible only to readers and institutions that can afford the pricey fees. (Individuals can subscribe to Nature for $199 per year or pay $8.99 per article, but university systems may pay as much as $11 million annually for a subscription to one of the big publishers’ lineup of journals.) Making discoveries accessible to anyone with an internet connection will level the playing field for individuals who lack a university affiliation, and for schools that can’t afford the costly library subscription fees.

But Springer Nature’s announcement also exposes a deep structural problem in scientific publishing. The proposed author fee, known as an article processing charge, or APC, is several times higher than what other publishers charge; it will likely be out of reach for researchers working outside of the world’s top institutions. Viewed in that light, Springer Nature’s move to open access seems less like a step toward equity and more like a corporation taking advantage of an uneven scientific funding landscape to increase its profits.

… Machine learning uses powerful algorithms to discover insights based on real-world data that can then be used to make predictions about future outcomes. As new data comes available, machine learning programs can automatically adapt and produce updated predictions. As with any tool, machine learning is not a silver bullet. However, there are many situations in which the technology can outperform linear and statistical algorithms.

Here are five of the most common use cases where machine learning can make a big difference:

1. When engineers can’t code rules for certain problems

After the Black Lives Matter protests last summer, big American employers vowed to improve racial diversity in their ranks. That included the tech industry, which has a lot of catching up to do.

Since last summer, Microsoft, Airbnb, Apple and Google have announced expansion plans or major investments in Atlanta. Such moves could make technology the leading driver of the city’s economic growth.

In Hobbs, New Mexico, the high school closed and football was cancelled, while just across the state line in Texas, students seemed to be living nearly normal lives. Here’s how pandemic school closures exact their emotional toll on young people.

One of the pivotal challenges of the next decade of artificial intelligence(AI) is to determine whether data and intelligence are democratized or remain in control of a few large organizations. A few months ago, I wrote a three-part series of the decentralization of artificial intelligence(AI). In that essay, I tried to cover the main elements that justify the movement of decentralized AI ranging from economic factors to technology enablers as well as the first generation of technologies that are developing decentralized AI platforms. The arguments made in those articles were fundamentally theoretical because, as we all know, the fact remains that AI today is completely centralized. However, as I work more in real-world AI problems, I am starting to realize that centralization is an aspect that is constantly hindering the progress of AI solutions. Furthermore, we should start seeing centralization in AI as a single problem but as many different challenges that surface at different stages of the lifecycle of an AI solution. Today, I would like to explore that idea in more detail.

What do I mean by claiming that AI has many centralization problems? If we visualize the traditional lifecycle of an AI solution we will see a cyclical graph that connects different stages such as model creation, training, regularization, etc. My thesis is that all those stages are conceptually decentralized activities that are boxed into centralized processes because of the limitation of today’s technologies.

The costs of anti-science sentiment in the United States have become magnified in the past year. The inability of scientists to convince the American public about the reality of COVID-19, the effectiveness of masks, and the safety of vaccines has led to loss of life and has decreased the possibility that the pandemic will end soon. We’ve seen the same problem of persuasion with the realities of climate change. Although many remedies have been proposed to combat this, the events of 2020 make clear that none of what we have been doing has worked. We need some new ideas. In an editorial this week, Aaron Mertz and Abhilash Mishra propose that one solution would be for the United States to launch an American Science Corps that would pay recent Ph.D.’s to fan out across the country and—armed with training in communications from appropriate experts—begin to build trust for science in the broader public.

There are a lot of exciting concepts in this proposal. They mainly focus on the need for the U.S. government to fund the corps. But there is another big challenge: getting the scientific community to value the kind of work that Mertz and Mishra are proposing. Scientists place a high priority on producing graduates who will go on to do more science. For years, we have known that only around 10% of science Ph.D.’s get tenure-track jobs, yet we describe the places where the other 90% end up as the “alternative careers.” In reality, the tenure-track job is the alternative career. Still, during my time in university administration, I was often visited by graduate students who were distraught because their adviser had lost interest in them after they revealed they wanted to work in industry. That broke my heart and also hurt my belief in the scientific enterprise. Shouldn’t we want our students to be doing what they find fulfilling?

“In terms of truly documented transmission between athletes during participation, I’m not aware of anything,” says Dr. Drew Watson, lead author on three University of Wisconsin studies that investigated COVID-19 risks in sports, and senior author on three other UW studies on the mental-health effects of sports shutdowns (all referenced below).

“I know researchers who are struggling to find even a single case among outdoor-sports participants, in particular.”

This year we significantly expanded the amount of data available in the report, worked with a broader set of external organizations to calibrate our data, and deepened our connections with Stanford HAI.

The 2021 report also shows the effects of COVID-19 on AI development from multiple perspectives. The Technical Performance chapter discusses how an AI startup used machine-learning-based techniques to accelerate COVID-related drug discovery during the pandemic, and our Economy chapter suggests that AI hiring and private investment were not significantly adversely influenced by the pandemic, as both grew during 2020. If anything, COVID-19 may have led to a higher number of people participating in AI research conferences, as the pandemic forced conferences to shift to virtual formats, which in turn led to significant spikes in attendance.

Last week, the NYPD sent a robotic dog to an active crime scene in the Bronx, where a home invasion was underway in a Wakefield area apartment building. The creepy footage of the bright-blue four-legged robot casually walking down the street raised our collective hackles. What exactly was it doing on the scene? And are we all about to enter a real-life episode of Black Mirror?

Predictably, the NYPD is tight-lipped about its newest robotic toy (as it is about virtually everything else it does). But here’s what we know so far.

Antenna, which says it uses data sampled from online bill payment services to assess what people are actually spending money on, has laid out the challenge facing ViacomCBS pretty clearly in the data sets below. But the easiest way of summing it up may be this way: (Just about) everyone already has Netflix.

This chart, for instance, tells us that 75 percent of people who newly got Netflix in the first half of 2020 are still paying for the service — a higher survival rate than all of its major streaming competitors. Meanwhile, only 34 percent of new 2020 Apple TV+ subscribers are still paying for the service now. (Antenna data does not include streamers who are getting free services from promotions like Disney’s Verizon bundle, or the free Apple TV+ trial Apple offers customers who buy some Apple hardware, like new iPhones.)

At the recent tinyML Summit 2021, Raspberry Pi co-founder Eben Upton teased the future of ‘Pi Silicon’ and it looks like machine learning could see a massive improvement thanks to Raspberry Pi’s news in-house chip development team.

It is safe to say that the Raspberry Pi Pico and its RP2040 SoC have been popular. The Pico has only been on the market for a few weeks, but already has sold 250,000 units with 750,000 on back order. There is a need for more boards powered by the RP2040 and partners such as Adafruit, Pimoroni, Adafruit and Sparkfun are releasing their own hardware, many with features not found on the Pico.

COVID-19 can be diagnosed in 55 minutes or less with the help of programmed magnetic nanobeads and a diagnostic tool that plugs into an off-the-shelf cellphone, according to Rice University engineers.

The Rice lab of mechanical engineer Peter Lillehoj has developed a stamp-sized microfluidic chip that measures the concentration of SARS-CoV-2 nucleocapsid (N) protein in blood serum from a standard finger prick. The nanobeads bind to SARS-CoV-2 N protein, a biomarker for COVID-19, in the chip and transport it to an electrochemical sensor that detects minute amounts of the biomarker.

Final results show 64.4% of voters coming out against the planned law on Sunday. The rejection rate among the cantons ranged between 70.7% and 55.8%.

At stake was the creation of the legal basis for a digital identity verification system, to be licenced and controlled by the state but provided mainly by private companies.

We now turn to understanding the impact that COVID-19 had on the personal productivity and well-being of information workers as their work practices were impacted by remote work. This chapter overviews people’s productivity, satisfaction, and work patterns, and shows that the challenges and benefits of remote work are closely linked. Looking forward, the infrastructure surrounding work will need to evolve to help people adapt to the challenges of remote and hybrid work.

Now that we have the academic research product track, are there any researchers/ students / R users interested in helping me add support for the v2 of the Twitter API including the new full-archive search in rTweet?