Data Science newsletter – November 9, 2021

Some students accumulated so many points early on that by the end of the term they knew they didn’t need to do more work and could still get an A. Others — often those who had to work or care for family members after school — would fail to turn in their homework and fall so far behind that they would just stop trying.

Do some types of information spread faster, broader, or further than others? To understand how information diffusions differ, scholars compare structural properties of the paths taken by content as it spreads through a network, studying so-called cascades. Commonly studied cascade properties include the reach, depth, breadth, and speed of propagation. Drawing conclusions from statistical differences in these properties can be challenging, as many properties are dependent. In this work, we demonstrate the essentiality of controlling for cascade sizes when studying structural differences between collections of cascades. We first revisit two datasets from notable recent studies of online diffusion that reported content-specific differences in cascade topology: an exhaustive corpus of Twitter cascades for verified true- or false-news content by Vosoughi et al. [S. Vosoughi, D. Roy, S. Aral. Science 359, 1146–1151 (2018)] and a comparison of Twitter cascades of videos, pictures, news, and petitions by Goel et al. [S. Goel, A. Anderson, J. Hofman, D. J. Watts. Manage. Sci. 62, 180–196 (2016)]. Using methods that control for joint cascade statistics, we find that for false- and true-news cascades, the reported structural differences can almost entirely be explained by false-news cascades being larger. For videos, images, news, and petitions, structural differences persist when controlling for size. Studying classical models of diffusion, we then give conditions under which differences in structural properties under different models do or do not reduce to differences in size. Our findings are consistent with the mechanisms underlying true- and false-news diffusion being quite similar, differing primarily in the basic infectiousness of their spreading process. [full text]

‘Perhaps the most surprising finding from our study is that, when we look at the overall picture, social factors are more important than environmental factors in explaining migration. And regardless of the level of income involved, gross national income was the key factor in explaining net-migration in half of countries,’ says Venla Niva, a doctoral student at Aalto University and lead author of the study published in Environmental Research Letters.

In their analysis, the team made use of a machine learning technique, called random forest analysis, well suited to dealing with the complex relationships seen between variables in very large sets of data. This allowed the researchers to explain the importance of each factor for each of the countries studied. Social factors were assessed in terms of income, education in years, life expectancy, government effectiveness; environmental factors were measured through natural hazards, water risk, food production scarcity, and drought prevalence.

At YouTube, our mission is to “give everyone a voice and show them the world.” Implicit in that mission is a sense of responsibility to our community: our users, our creators, and our advertisers. Responsibility is our number one priority at YouTube, and we want to protect our community while enabling new and diverse voices to break through.

The Government of Canada is drafting legislation to address “Harmful Content Online” and we are committed to helping them achieve that objective. Everyone deserves to feel safe online. At YouTube we feel a deep responsibility to keep our users safe and remove content that violates our policies. Part of that responsibility includes working together with governments and other stakeholders to get regulatory frameworks right.

Geoffrey Challen, associate professor in engineering, created a course website as a resource for students learning how to code. Now, he leads a new project where students are creating a communal website in which they can add and learn from valuable information.

The CS 124: Introduction to Computer Science I website underwent big changes during the pandemic to support students in online learning.

“(Some of the changes include) the new daily lesson format, containing the interactive walk-throughs, new homework problems that were completed through the website, a new quiz system and a new online help site where students can ask questions and get assistance from staff with their code,” Challen said.

Announcing a new university dedicated to the fearless pursuit of truth: @uaustinorg

Today will I take you through the keyhole to look at a group of software systems not well known to the public, which I call “Bank Python”. Bank Python implementations are effectively proprietary forks of the entire Python ecosystem which are in use at many (but not all) of the biggest investment banks. Bank Python differs considerably from the common, or garden-variety Python that most people know and love (or hate).

Thousands of people work on – or rather, inside – these systems but there is not a lot about them on the public web. When I’ve tried to explain Bank Python in conversations people have often dismissed what I’ve said as the ravings of a swivel-eyed loon. It all just sounds too bonkers.

As the first large initiative inspired by Northeastern’s new academic plan, Experience Unleashed, the university plans to hire 500 new faculty members across its global network over the next five years. New arrivals will be full-time research faculty oriented toward developing interdisciplinary, collaborative solutions for the world’s greatest challenges. Photo by Matthew Modoono/Northeastern University

The NIH’s pre-grant peer-review process requires that many reviewers approve an application. This consensus-oriented style can be a check against novelty—what if one scientist sees extraordinary promise in a wacky idea but the rest of the board sees only its wackiness? The sheer amount of work required to get a grant also penalizes radical creativity. Many scientists, anticipating the turgidity and conservatism of the NIH’s approval system, apply for projects that they anticipate will appeal to the board rather than pour their energies into a truly new idea that, after a 500-day waiting period, might get rejected. This is happening in an academic industry where securing NIH funding can be make-or-break: Since the 1960s, doctoral programs have gotten longer and longer, while the share of Ph.D. holders getting tenure has declined by 40 percent.

Fast Grants aimed to solve the speed problem in several ways. Its application process was designed to take half an hour, and many funding decisions were made within a few days. This wasn’t business as usual. It was Operation Warp Speed for science.

There is no question that equipping labs with information technology is a rapidly changing endeavor given the now-ubiquitous use of cloud computing and the soon-to-be ubiquitous deployment of artificial intelligence and machine learning. And the velocity of change in drug discovery IT is increased by something advancing even more rapidly—science.

New technologies such as gene editing and genomics place drug discovery on the front line of a transformation in data management. As laboratories produce petabytes of data, responsibility for managing and curating information in the hunt for new drugs is increasingly shifting from centralized IT departments to research teams at the bench. It’s a big change for both the scientist and the IT department.

“It is absolutely clear that the rate of scientific innovation far outstrips the speed of IT innovation. It also moves faster than I, an IT person, can build a storage system or build an analytical environment or build a database,” says Chris Dagdigian, senior director of BioTeam, a life sciences IT consultancy.

Chinese drug discovery biotech HitGen is teaming up with U.K.-based Cambridge Molecular to better seek out new drugs.

The pair has penned an exclusive alliance, financials of which were not shared, to create the so-called DeepDELve 2 – Cambridge Molecular’s highly optimized DNA encoded library (DEL)-specific deep learning system, which will be wedded to HitGen’s own DEL discovery platform.

DeepDELve 2 is a “robust and highly specialised deep learning pipeline,” according to the partners, that allows clients to readily access potential ligands for drug targets.

Google parent company Alphabet has launched a new drug discovery company in the U.K. called Isomorphic Labs.

The company will build on research carried out by London artificial intelligence lab DeepMind, which Google acquired in 2014. While the firm was only officially announced on Thursday, it was incorporated in February, according to a filing with Companies House, the U.K. company registry.

Two schools at the University of Virginia this week announced a unique partnership to explore ways to combine the power of data science with the teaching and practical applications of business.

The School of Data Science and the Darden Graduate School of Business “collaboratory” capitalizes on the explosion of data; the ability to analyze and pull insights from it; and the shared interest in applying that new knowledge to improving how business is taught, researched and applied in actual settings.

The new Collaboratory for Applied Data Science in Business, with the support of the UVA Office of the Executive Vice President and Provost, will build on the existing relationship between the two schools and provide new avenues for students, staff and faculty to work and learn together at the intersection of data science and business.

Cutting-edge quantum computing will one day run Sydney’s vast transport network under a world-first plan to use the technology that experts say can solve complex problems in seconds, rather than centuries.

The NSW government is set to brief the technology industry in coming weeks about a plan to establish a quantum technology hub near Central Station to run the city’s transport network, with contracts set to be awarded in 2022.

At the age of 98, former Secretary of State Henry Kissinger has a whole new area of interest: artificial intelligence. He became intrigued after being persuaded by Eric Schmidt, who was then the executive chairman of Google, to attend a lecture on the topic while at the Bilderberg conference in 2016. The two have teamed up with the dean of the MIT Schwarzman College of Computing, Daniel Huttenlocher, to write a bracing new book, The Age of AI, about the implications of the rapid rise and deployment of artificial intelligence, which they say “augurs a revolution in human affairs.” The book argues that artificial intelligence processes have become so powerful, so seamlessly enmeshed in human affairs, and so unpredictable, that without some forethought and management, the kind of “epoch-making transformations” they will deliver may send human history in a dangerous direction.

In late 2019 we started to work on the concept of AURA – the Ada User Repository Annex. Our goal was simple, but specific: to build a native package manager and build system for Ada.

CRA-I held its second roundtable focused on Best Practices on using the Cloud for Computing Research. The purpose of this roundtable was to discuss best practices and the resulting synergistic opportunities across industry, academia, and government. See a video of the roundtable here. The session was moderated by two members of the CRA-I steering committee: Fatma Ozcan (Google) and CRA-I Co-Chair Vivek Sarkar (Georgia Institute of Technology). The panelists were David Culler (Google), Ed Lazowska (University of Washington), Margaret Martonosi (National Science Foundation), Giovanni Pacifici (IBM Research), and Raghu Ramakrishnan (Microsoft).

Overwhelmingly, the panelists agreed that using the commercial cloud for computing research makes you, as Pacifici put it, “more productive, gives you access to the latest technology, and provides agility.”