|
|
|
|
|
Pittsburgh Hospital Maps Infections; High-Frequency Traders Face Cloudier Futures
|
|
|
|
|
|
Welcome back. The issue seems simple enough: People go to hospitals when they’re sick, and sick people can spread viral infections. But even with wearing masks and disinfecting hospital beds, stethoscopes and other medical equipment, efforts to prevent hospital infections continue to come up short. About one in 31 U.S. patients contracts at least one infection from a hospital stay, according to the Centers for Disease Control and Prevention.
The Covid-19 pandemic has made the issue even more pressing, by both crowding hospitals and raising the stakes of catching a disease during a trip to the ER—whether for Covid-19 symptoms or a twisted ankle.
Doctors at the University of Pittsburgh Medical Center think they may have a solution. Researchers at the center’s flagship hospital are testing an AI-powered system designed to track the path of hospital-related infections. The system is set up to crunch a giant dataset combining bacterial data, infection cases and electronic health records of where patients were in the hospital and what procedures they had, among other variables. The goal is to identify hazardous situations and stop viruses before they spread.
Had the system been deployed at the hospital over the past two years, early tests suggest, it could have heloed prevent dozens of viral infections, including Covid-19.
|
|
|
|
|
Dr. Lee Harrison, left, a professor at the University of Pittsburgh School of Medicine, led the team of Pitt and Carnegie Mellon University researchers who developed an AI system. PHOTO: NATHAN LANGER, UPMC
|
|
|
|
Tracking Hospital Infections
|
|
Doctors think a new AI-powered system may help them map an infection's path through a hospital and take the right steps to stop the spread, including enhanced measures to disinfect rooms and sterilize equipment, WSJ’s John McCormick reports.
What it can do. At its flagship hospital, the University of Pittsburgh Medical Center says it is deploying a machine-learning system designed to help spot the beginnings of hospital-based infectious disease outbreaks.
How it works. The system uses machine learning to detect patterns in a combination of bacterial DNA sequencing data, infection cases and electronic health record medical records of where patients stayed in the hospital, the procedures they’ve had, the equipment used to treat them, and the health workers with which they’ve had contact.
Is it effective? Based on two-year’s worth of stored samples and historical patient data, researchers say the system could have prevented more than 60 transmissions of infectious diseases at UPMC Presbyterian.
|
|
|
|
The Pentagon said Friday that it has issued formal solicitations to Amazon.com Inc., Microsoft Corp., Alphabet Inc.’s Google and Oracle Corp. for a multibillion-dollar cloud-computing contract to replace its scrapped JEDI procurement, which became mired in litigation and a torrent of objections from Congress, The Wall Street Journal reports.
|
|
|
The Pentagon. PHOTO: STEFANI REYNOLDS/BLOOMBERG NEWS
|
|
|
New deal, same goals. Like the original JEDI contract, which was valued at up to $10 billion over 10 years, the new initiative is aimed at putting the Defense Department on a stronger footing to develop AI capabilities that are seen as vital to military operations in the future.
|
|
|
Second chances. After losing its bid to Microsoft, Amazon later complained in court that the original award to Microsoft was the result of political meddling by then-President Donald Trump, a charge the Pentagon denied.
Back in the game. Google, which came under fire from employee activists for working with the miliary and dropped out of the JEDI bid process, said it would proudly work to help modernize the Pentagon's operations under the new contract.
|
|
|
|
|
1 in 31
|
The ratio of U.S. patients that contract at least one infection during a hospital stay, according to the Centers for Disease Control and Prevention.
|
|
|
|
|
|
|
|
CME Group offices in New York. PHOTO: BRENDAN MCDERMID/REUTERS
|
|
|
|
Google, CME Partnership Could Upend Tech Investments by Traders
|
|
CME Group Inc.’s plan to move core trading systems to Google Cloud, aimed at tapping AI and other advanced software tools, could upend hundreds of millions of dollars in investments by traders in communications and data centers, industry analysts tell The Wall Street Journal.
What’s new. CME and Alphabet Inc.’s Google this month announced a 10-year agreement to move CME applications to the cloud. The applications will move to the cloud in phases, with core trading systems eventually making the shift.
Why the change. CME said the deal will allow it to bring on new users faster, streamline operations and develop new tools with Google technology, such as AI software for monitoring market risks.
Why it matters. Citing regulatory and security issues, high-frequency traders have spent years avoiding third-party cloud providers, opting instead to build their own state-of-the art systems geared for extremely time-sensitive markets.
|
|
|
|
“This is a harder than usual cloud migration.”
|
— Philip Moyer, Google Cloud's vice president of strategic industries, on efforts to move CME Group's trading systems to the cloud
|
|
|
|
|
|
WSJ Pro AI Extra: Self-Checkout Tech
|
|
|
|
There’s no question that self-checkout has improved over the years. Shoppers today are a lot less likely to get an “unexpected item in the bagging area,” forcing them to wait for a store employee to help. But the customer experience is still far from perfect. Here are a few high-tech solutions:
-
Toshiba Corp, which makes self-checkout kiosks, says it is working with retailers on computer-vision and AI-based tools to allow machines to more easily recognize difficult-to-scan items.
-
NCR Corp. offers software for self-checkout machines that uses computer-vision technology to identify produce that is placed on the scale. It then brings up a shortlist of the most likely items from which the customer can choose.
-
Amazon.com Inc.’s Just Walk Out technology automatically detects when products are taken from or returned to the shelves, and it keeps track of them in a virtual cart. When customers are finished shopping, they skip the checkout line and leave the store and are automatically charged.
|
|
|
|
|
Netomi Chief Executive Puneet Mehta. PHOTO: NETOMI INC.
|
|
|
|
Chatbot startup fetches higher valuation. Netomi Inc., a provider of customer-service chatbots, said it has raised $30 million in a Series B funding round that values the company at $210 million, up from a valuation of $67.5 million in 2019. (The Wall Street Journal)
Ericsson to acquire Vonage. Ericsson AB agreed to buy cloud-communications provider Vonage Holdings Corp. for $6.2 billion, a deal that the Swedish telecom-equipment maker said will expand its wireless enterprise business with Vonage-developed AI chatbots and other tools. (The Wall Street Journal)
Chamber seeks AI records. As part of a series of letters and Freedom of Information Act requests related to Federal Trade Commission antitrust efforts, the U.S. Chamber of Commerce says it wants to see “all records related to artificial intelligence” created during the Biden administration. (The Wall Street Journal)
Australia to beef up tech. Under a new security partnership with the U.S., the Australian government says it will strengthen the county’s capabilities in AI, quantum computing and other technologies that could be used to counter threats from China. (The Wall Street Journal)
Less code, smarter apps. By simplifying the process of building software applications, tech insiders say low- and no-code platforms are poised to accelerate the spread of AI-driven capabilities at companies across the economy (ZDNet)
Lie detector catches 'micro expressions.' Researchers at Tel Aviv University are developing an AI-enabled tool designed to read facial muscles that appear for less than 60 milliseconds when a test subject is lying. (WION)
NYC to rein in high-tech hiring tools. A New York City council bill approved earlier this month would ban employers from using automated hiring tools, unless they conduct bias audits to show algorithms aren’t discriminating based on an applicant’s race or gender. (The Associated Press)
Startup wins Space Force contract. Cognitive Space, a Houston-based AI startup that recently closed a $4 million seed funding round, won a $1.7 million contract to provide the federal space agency with AI-enabled software designed to manage imaging satellite constellations. (SpaceNews)
-
New space race. Space-data companies, which use satellites to snap photos of Earth, track radio signals and use radar to peer through clouds, raised $5.2 billion last year, up from $1.4 billion in 2015, according to data from PitchBook. (The Wall Street Journal)
Roommates launch real-estate platform. Zuma, an AI-powered virtual leasing office developed during the Covid-19 lockdown by roommates Shiv Gettu and Kendrick Bradley, launched last week with $6.7 million in a seed-funding round led by Andreessen Horowitz. (Forbes)
|
|
|
Samsung Electronics Co. plans to build a roughly $17 billion chip-making plant in Taylor, Texas, according to people familiar with the matter, a mega investment by the South Korean tech giant, as the Biden administration pushes for an expansion of U.S. semiconductor production. (The Wall Street Journal)
Tesla Inc. plans to launch its Model S Plaid sedan in China next year, "probably around March," Chief Executive Elon Musk said in a tweet Monday. (The Wall Street Journal)
Comcast Corp.’s NBCUniversal is considering removing much of its content from Hulu and making it exclusive to its Peacock platform, according to people familiar with the matter, as the media giant determines how to best play its hand in the streaming wars. (The Wall Street Journal)
Activision Blizzard Inc. Chief Executive Bobby Kotick has told senior managers he would consider leaving the company if he can’t quickly fix the culture problems at the videogame giant, according to people familiar with his comments. (The Wall Street Journal)
|
|
|
|
|
|