Bloomberg Technology; By Chiara Albanese , Thomas Seal , and Rodrigo Orihuela
from
Shortly after coronavirus forced Italian Prime Minister Giuseppe Conte to lock down the country, lawmaker Massimiliano Capitanio took an unusual call at his office in Rome.
It was an appeal for help from a hospital at the epicenter of the outbreak in northern Italy. Its administrators direly needed faster internet connections and computers to deal with the flood of patients. Capitanio — who sits on parliament’s telecommunications committee — called the country’s phone companies to help out.
To Capitanio, the pandemic was a wake-up call to fix Italy’s creaking internet. Now Conte has stepped in with a plan to kick-start investment by merging the country’s two biggest landline networks.
Data science has primarily focused on big data, but for many physics, chemistry, and engineering applications, data are often small, correlated and, thus, low dimensional, and sourced from both computations and experiments with various levels of noise. Typical statistics and machine learning methods do not work for these cases. Expert knowledge is essential, but a systematic framework for incorporating it into physics-based models under uncertainty is lacking. Here, we develop a mathematical and computational framework for probabilistic artificial intelligence (AI)–based predictive modeling combining data, expert knowledge, multiscale models, and information theory through uncertainty quantification and probabilistic graphical models (PGMs). We apply PGMs to chemistry specifically and develop predictive guarantees for PGMs generally. Our proposed framework, combining AI and uncertainty quantification, provides explainable results leading to correctable and, eventually, trustworthy models. The proposed framework is demonstrated on a microkinetic model of the oxygen reduction reaction. [full text]
In a paper posted online in July, [Nathan] Klein and his advisers at the University of Washington, Anna Karlin and Shayan Oveis Gharan, have finally achieved a goal computer scientists have pursued for nearly half a century: a better way to find approximate solutions to the traveling salesperson problem.
This optimization problem, which seeks the shortest (or least expensive) round trip through a collection of cities, has applications ranging from DNA sequencing to ride-sharing logistics. Over the decades, it has inspired many of the most fundamental advances in computer science, helping to illuminate the power of techniques such as linear programming. But researchers have yet to fully explore its possibilities—and not for want of trying.
State of Kentucky, Cabinet for Health and Family Services, Office of Health Data and Analytics
from
The State University Partnership (SUP) is a research program between Cabinet for Health and Family Services (CHFS) leadership and university research centers designed to expand Cabinet research capacity, support development of evidence-based policy and foster a research community in the commonwealth.
The White House Office of Management and Budget (OMB) said Friday that federal agencies will use artificial intelligence to eliminate outdated, obsolete, and inconsistent requirements across tens of thousands of pages of government regulations.
A 2019 pilot project used machine learning algorithms and natural language processing at the Department of Health and Human Services. The test run found hundreds of technical errors and outdated requirements in agency rulebooks, including requests to submit materials by fax.
OMB said all federal agencies are being encouraged to update regulations using AI and several agencies have already agreed to do so.
Speaking at a POLITICO AI Summit on Thursday, Eric Schmidt, chairman of the National Security Commission on Artificial Intelligence and former CEO at Google, said the U.S. urgently needs a national AI strategy based on the principle of “whatever it takes.” Schmidt said Americans could not relax on AI issues because even consumer AI innovations have the potential to be “used for cyber war” in ways that aren’t always evident or anticipated. Schmidt has previously warned against “high tech authoritarianism.”
While the U.S. has lacked central organizing of its AI, it has an advantage in its flexible tech industry, said Nand Mulchandani, the acting director of the U.S. Department of Defense Joint Artificial Intelligence Center. Mulchandani is skeptical of China’s efforts at “civil-military fusion,” saying that governments are rarely able to direct early stage technology development.
ICML had the Machine Learning Journal; NIPS had Neural Computation. But then ICML types like me started attending NIPS (and so did computational statisticians such as Leo Breiman and Jerry Friedman). Learning theory was at both meetings + COLT
Recently, CDS PhD student and Deepmind Fellow, William Falcon recently raised an $18.6 Series A for his latest entrepreneurial endeavor, Grid AI. CDS social media team member, Colton Laferriere, spoke to William about his background, his entrepreneurship and how to get your ideas off the ground.
Cornell researchers have invented an earphone that can continuously track full facial expressions by observing the contour of the cheeks – and can then translate expressions into emojis or silent speech commands.
With the ear-mounted device, called C-Face, users could express emotions to online collaborators without holding cameras in front of their faces – an especially useful communication tool as much of the world engages in remote work or learning.
Before the coronavirus pandemic, NearForm, an enterprise software company headquartered in the tiny southeastern Irish seaside town of Tramore, helped some of the most recognizable brands in the world—from Uber to Condé Nast—make quick digital transformations.
When COVID-19 hit and the Irish government needed a technology company with a reputation for being quick and nimble to build a contact tracing app in March, NearForm jumped at the chance to help.
The company’s software engineers, who were all used to working remotely even before the pandemic, created COVID Tracker, a decentralized app that keeps users anonymous but alerts them if they’ve crossed paths with someone who has tested positive for coronavirus in the past two weeks. With a 35% adoption rate, it has been such a success story in Ireland that versions are now being used in four U.S. states, with many more expressing interest.
The University of Calgary is adding a new graduate program offering to help both new and experienced professionals transition to a career in tech.
The Master in Data Science and Analytics (MDSA) is a multidisciplinary, professionally oriented graduate degree program offered through a collaboration between the Faculty of Science, the Haskayne School of Business, the Cumming School of Medicine, and the Faculty of Graduate Studies.
In line with the University of Calgary’s Unstoppable “growth through focus” plan to create Canada’s most entrepreneurial university by expanding graduate education geared toward professionals and strengthening ties to industry, the MDSA will provide students who have backgrounds in a variety of different disciplines with the skills and abilities needed to become data scientists or data analysts, and to flourish in a global economy that is increasingly data-driven.
“The long-term vision is this concept of programmable immunity,” said Kwong, associate professor in the Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University, who partnered with fellow researcher Brandon Holt on the paper, “Protease circuits for processing biological information,” published Oct. 6 in the journal Nature Communications. The research was sponsored by the National Institutes of Health.
The story of this paper begins two years ago when, Holt said, “our lab has a rich history of developing enzyme-based diagnostics; eventually we started thinking about these systems as computers, which led us to design simple logic gates, such as AND gates and OR gates. This project grew organically and we realized that there were other devices we can build, like comparators and analog-digital convertors. Eventually this led to the idea of taking an analog-to-digital converter and using that to digitize bacterial activity.”
Ultimately, they assembled cell-free bio-circuits that can combine with bacteria-infected blood, “with the basic idea that it would quantify the bacterial infection — the number of bacteria — then calculate and release a selective drug dose, essentially in real time,” said Holt, a Ph.D. student in Kwong’s Laboratory for Synthetic Immunity and lead author of the paper.
Harvard University, Harvard T.H. Chan School of Public Health, The Forum
from
Online October 21, starting at 1 p.m. “With Ed Yong, Science Writer, The Atlantic; Moderated by Meg Tirrell, Senior Health and Science Reporter, CNBC” [registration required for Zoom webinar]
“This competition uses NFL’s Next Gen Stats data, which includes the position and speed of every player on the field during each play. You’ll employ player tracking data for all drop-back pass plays from the 2018 regular season. The goal of submissions is to identify unique and impactful approaches to measure defensive performance on these plays.” Deadline for submissions is January 7, 2021.
This week at the 2020 International Society for Music Information Retrieval Conference, Spotify open-sourced Klio, an ecosystem that allows data scientists to process audio files (or any binary files) easily and at scale. It was built to run Spotify’s large-scale audio intelligence systems and is leveraged by the company’s engineers and audio scientists to help develop and deploy next-generation audio algorithms.
This week, we talked to Joe Hellerstein, a computer science professor at the University of California, Berkeley, who created one of the most widely used tools for preparing data for machines. Joe talked about the challenges of data wrangling and how the field is evolving in the age of AI. [audio, 44:13]
arXiv readers now have a faster way to find articles relevant to their interests. From an article abstract page, readers can simply activate the CORE Recommender to find additional open access research on similar topics.
The Recommender, part of the arXivLabs toolset, was developed by CORE, a global aggregator of open access scientific content, which provides access to millions of full texts. CORE’s mission is to aggregate all open access research outputs from repositories and journals worldwide and make them available to the public. In this way, CORE facilitates free unrestricted access to research for all.
