Data Science newsletter – June 25, 2021

Peter Scholze wants to rebuild much of modern mathematics, starting from one of its cornerstones. Now, he has received validation for a proof at the heart of his quest from an unlikely source: a computer.

Although most mathematicians doubt that machines will replace the creative aspects of their profession anytime soon, some acknowledge that technology will have an increasingly important role in their research — and this particular feat could be a turning point towards its acceptance.

Scholze, a number theorist, set forth the ambitious plan — which he co-created with his collaborator Dustin Clausen from the University of Copenhagen — in a series of lectures in 2019 at the University of Bonn, Germany, where he is based. The two researchers dubbed it ‘condensed mathematics’, and they say it promises to bring new insights and connections between fields ranging from geometry to number theory.

In “Power of data in quantum machine learning”, published in Nature Communications, we dissect the problem of quantum advantage in machine learning to better understand when it will apply. We show how the complexity of a problem formally changes with the availability of data, and how this sometimes has the power to elevate classical learning models to be competitive with quantum algorithms. We then develop a practical method for screening when there may be a quantum advantage for a chosen set of data embeddings in the context of kernel methods. We use the insights from the screening method and learning bounds to introduce a novel method that projects select aspects of feature maps from a quantum computer back into classical space. This enables us to imbue the quantum approach with additional insights from classical machine learning that shows the best empirical separation in quantum learning advantages to date.

2. This thread describes the paper and the backstory leading to it. I’ll be posting over the course of the day as I can find the time.

Three years ago, @uwcip
postdoc @jbakcoleman
organized a summer meeting at Princeton. I attended, and it changed the direction of my research.

.

.

.
31. We don’t have the theory in place to understand how the design of social networks drives network structure and function and how that in turn drives the flow of information or disinformation through communication networks. Yet we can’t wait to figure it out.

32. We need to be acting and learning in parallel.

We also need all hands on deck. Crisis disciplines are radically transdisciplinary. We need to integrate a breadth of perspectives and we cannot predict from where the crucial insights will emerge to solve the problems we face.

A recent study revealed that as a city becomes more economically developed, its citizens’ happiness becomes more directly related to the area of urban green space.

A joint research project by Professor Meeyoung Cha of the School of Computing and her collaborators studied the relationship between green space and citizen happiness by analyzing big data from satellite images of 60 different countries.

A brand new institute at the University of Waterloo will harness the power of academia to address social, environmental, and economic challenges threatening the sustainability of aviation and aerospace.

The Waterloo Institute for Sustainable Aeronautics (WISA) – officially approved by the university Senate on June 21 – will serve as a portal for industry stakeholders to access the university’s extensive knowledge network. The new organization will be led by Dr. Suzanne Kearns, associate professor of aviation at the University of Waterloo in the Faculty of Environment.

Think of them as master Lego builders, only at an atomic scale. Engineers at CU Boulder have taken a major step forward in combing advanced computer simulations with artificial intelligence to try to predict how electronics, like the transistors in your cell phone, will fail.

The new research was led by physicist and aerospace engineer Sanghamitra Neogi and appears this week in the journal npj Computational Materials.

In their latest study, Neogi and her colleagues mapped out the physics of small building blocks made up of atoms, then used machine learning techniques to estimate how larger structures created from those same building blocks might behave. It’s a bit like looking at a single Lego brick to try to predict the strength of a much larger castle.

For urban planners, parking rules established decades ago have become a contentious 21st-century challenge. Parking takes up about one-third of land area in U.S. cities; nationwide, there are an estimated eight parking spaces for every car.

In 2017, Buffalo, New York, became the first U.S. city to stop requiring development projects to include at least a minimum amount of parking. Other cities followed, including Hartford, Connecticut, and Santa Monica, California. Many cities are now considering reforms, and a bill pending before the California Legislature would remove minimums for new buildings near public transportation across the Golden State.

But despite growing support for parking reform, there is little data showing how such changes affect urban development.

Many streamers regularly take advantage of playlist features intended to highlight new music. Half say that personalized playlists do a good job of exposing them to new music; 82% like the mix of old and new music they are hearing. Playing similar songs after a listener selects something was a top-rated feature for aiding discovery. Innovation in streaming features isn’t just engineering for engineering sake — fans are embracing these features in their discovery journeys.

For years, the conversation was about which music format was No. 1 when it came to music discovery. In retrospect that was the wrong conversation. The thinking should be about how to serve fans and help discovery to thrive. Today, streaming is the most influential source, but there are a host of other touchpoints that can help artists to promote their music and help fans to find it.

Prior to the COVID-19 pandemic, digital products were being adopted at ~34% CAGR in research studies (Marra et al., 2020), which has further accelerated over the past year and a half. With this pivotal shift toward collecting digital measures using connected sensors, it is vital to evaluate their accuracy and determine how measurement errors may impact research conclusions and healthcare decision-making.

The collaboration will leverage HumanFirst’s Atlas platform and DCRI’s deep expertise in trial design to create and conduct innovative and scientifically rigorous protocols to evaluate sensors and other digital measures on behalf of sponsor partners. The new Digital Measures Evaluation Center will perform fit-for-purpose technical and clinical evaluations that are designed to align with sponsor needs.

Are you more likely to discover your next favorite song on the radio or through one of many playlist’s put together for you like Spotify’s ‘Spotify Weekly’? We talk with a media professor about how artificial intelligence and technology is changing how we discover music. [audio, 21:38]

DOD is creating operational data teams that will be dispatched to all 11 combatant commands. The teams will rapidly work, catalog, manage and automate data feeds that inform decision making. These teams will remain to ensure data is captured, complete, curated and usable until combatant commands can leverage the data needed to create decision advantage.

DOD will build on new sustained data relationships with additional “flyaway teams of technical experts” to help combatant commands streamline and automate workflows through the integration of AI. The teams will bring top-tier talent and technology, building real capabilities that can be evaluated in real operational environments.

DOD will use the information gathered from the data teams, the AI flyaway teams and combatant command exercises to update network infrastructure, remove policy barriers and ensure the reliability and effectiveness of its global warfighting capabilities.

“I think the numbers tell us that, all in all, our institutions are doing a pretty good job of trying to mitigate the number of students who actually get sent to debt collectors,” said Nate Mackinnon, executive director of the Massachusetts Association of Community Colleges. “Getting to the point of sending something to a debt collector is obviously our path of last resort.”

He said the state’s comparatively low level of financial support for public universities and colleges ties their hands.

“Ideally, we’d send zero students to debt collectors,” Mackinnon said. “Unfortunately, we are a high-cost state when it comes to community colleges and public education. We don’t enjoy the state’s support at the level that other states do.”

Determined AI, a machine learning technology company co-founded by Ameet Talwalkar, an assistant professor in the Machine Learning Department at Carnegie Mellon University’s School of Computer Science, will join Hewlett Packard Enterprise (HPE). …
The founders wrote that HPE will continue to expand Determined AI’s training platform as an open-source project. The platform allows engineers to easily implement and train machine learning models to provide faster and more accurate insights from data in almost every industry. For example, Determined AI’s platform sped up the training of a machine learning model for drug discovery from three days to three hours.

Iowa became home Tuesday to the National Science Foundation’s fourth and final wireless research platform, thanks to a $16 million investment in Iowa State University-led research aimed at achieving universal and affordable rural broadband.

ISU joins New York City, Salt Lake City and North Carolina State University as a research platform host for the NSF-backed “Platforms for Advanced Wireless Research” program, a $100 million public-private partnership designed to speed “development and commercialization of promising technologies and applications.”

The ISU-based platform has been labeled, “ARA: Wireless Living Lab for Smart and Connected Rural Communities,” and complements the other platforms by adding a focus on technologies to improve rural broadband connectivity.

The algorithms underpinning artificial intelligence and machine learning increasingly influence our daily lives. They can decide everything from which video we’re recommended to watch next on YouTube to who should be arrested based on facial recognition software.

But the data used to train these systems often replicate the harmful social biases of the engineers who build them. Eliminating this bias from technology is the focus of Code^Shift, a new data science lab at Texas A&M University that brings together faculty members and researchers from a variety of disciplines across campus.

It’s an increasingly critical initiative, said Lab Director Srividya Ramasubramanian, as more of the world becomes automated. Machines, rather than humans, are making many of the decisions around us, including some that are high-risk.

In January 2021, we launched ‘A New AI Lexicon:’ a call for contributions to generate alternate narratives, positionalities, and understandings to the better known and widely circulated ways of talking about AI. In our introductory call, we identified the contours and politics of the critical AI space, as well as the silences and erasures they end up producing. We were hopeful yet uncertain about whether or who this call would speak to. We began a process of reimagining AI futures by rearticulating the scope and boundaries of critical AI. We are by no means the only group of scholars looking for and pushing the boundaries of AI talk and buzzwords, and are grateful that so many activists, practitioners, and scholars joined our call. We received over 130 pitches for contributions and are truly grateful for and overwhelmed by the interest. After several months of communing and conversing with the authors, we are thrilled to publish the first set of essays as a part of ‘A New AI Lexicon’ starting this week.

From the submissions we received, we selected over 40 essays, all centered around keywords new and old. As we publish the final essays, we hope to show how multiple essays by different authors convene under the same keyword but speak to it in vastly different ways; how familiar buzzwords in the discourse can be reinterpreted and re-tooled towards different ends; and how new frames can redirect our attention towards underexplored contexts and geographies.