|
Covid-19 Unmasks Tech Strengths and Weaknesses; Investors Back Smarter Robots
|
|
|
|
|
|
Good morning. The recent upswing in coronavirus cases is set to trigger another round of lockdowns and travel restrictions that have kept workers in their homes. For big cloud providers, which increasingly are embedding smart software into enterprise applications, surging demand for remote-work tools has lifted sales to record highs. But even as AI spreads across business sectors, its ability to supercharge efforts to combat the crisis has fallen far below expectations, revealing limits to a nascent technology that’s still finding its legs.
|
|
|
|
|
Carolina Barcenas of Visa and Gary Marcus of Robust AI speaking at the WSJ Pro AI executive forum on Wednesday. PHOTO: THE WALL STREET JOURNAL
|
|
|
|
When the coronavirus outbreak struck, scientists and medical researchers had good reason to believe AI would help accelerate the search for a cure. Its failure to do so stems from a number of shortcomings, robotic software developer Gary Marcus told attendees last week at the WSJ Pro Artificial Intelligence Executive Forum.
Lack of data. “With something like Covid, it’s difficult because we don’t have all the information we need,” said Mr. Marcus, chief executive of Robust AI Inc.
Lack of ability. Smart software today doesn’t have the intuitive capabilities needed to read and fully understand the mounting piles of medical literature, he said.
Lack of privacy. Mr. Marcus said data privacy and other ethical concerns can impede the use of AI in areas like contact tracing.
|
|
|
Amazon, Google and other enterprise tech providers are posting record earnings, as the crisis boosts demand for cloud-based remote-work capabilities, including AI-enabled business apps, The Wall Street Journal reports.
|
|
AWS delivers. Amazon’s cloud unit helped propel the company's third-quarter revenue to an all-time high of $96.2 billion, tripling profit to $6.3 billion, on a sharp rise in online retail, social-media and cloud services.
|
|
|
Google Alert. After a slump earlier in the year, Google parent Alphabet Inc. reported third-quarter profit of $11.2 billion, with revenue from Google Cloud climbing 44%, to $3 billion for the quarter.
|
|
|
|
$22.5 million
|
Total funding raised by robotic software startup Robust AI Inc., including a $15 million round announced last week at The Wall Street Journal’s WSJ Pro Artificial Intelligence Executive Forum
|
|
|
|
|
|
Sebastian Thrun, CEO of Kitty Hawk and chairman and co-founder of Udacity, at the WSJ Pro Artificial Intelligence Executive Forum. PHOTO: THE WALL STREET JOURNAL
|
|
|
Where the Rubber Meets the Road
|
|
Sebastian Thrun, co-founder of online-learning platform Udacity Inc. and CEO of flying-car company Kitty Hawk Corp., shared advice for tech innovators at The Wall Street Journal’s WSJ Pro AI Executive Forum.
Who knows. “You can’t expect to know everything on day one. You have to admit that you don’t know, and that means you will make choices that are poor choices in hindsight.”
Just dig it. “The technology is the shovel you use to dig for gold, and it’s the gold you’re after, not the shovel. And that’s where some people fail.”
|
|
|
|
"We often go in and say, OK, let’s try a thing and see where this ends up. And very frequently, we come to a dead end and say, that didn’t work."
|
— Sebastian Thrun, CEO of flying-car company Kitty Hawk Corp.
|
|
|
|
|
|
Online retailers like Wayfair, Etsy and Pinterest aim to use e-commerce data to improve their search-and-recommendation engines. PHOTO: KEVIN HAND
|
|
|
Retailers use AI to improve online recommendations for shoppers. Since the coronavirus outbreak, online retailers are ratcheting up efforts to leverage data from a surge in e-commerce to get better at helping customers find what they are looking for. Two AI technologies—computer vision and natural-language processing—are helping stores get better at building predictive models with a level of accuracy unimaginable just a few years ago.
Computer vision. A computer vision algorithm from retailer Wayfair Inc. captures design features of products and then takes that information, along with tags added by professional designers, and applies it to similar items in the store’s extensive line of products. The richer the list of keywords tagged to a product, the more likely an algorithm is to produce relevant search results. “You’re taking a very large catalog and parsing it and segregating it into searchable terms, to match what you want,” Wayfair CTO Jim Miller tells The Wall Street Journal.
Natural-language processing. Etsy Inc.’s search engines learn from a customer’s past queries as well as from other customers’ past queries and purchases. Billions of historical data points are consulted to make links between the terms shoppers have used and the products they eventually find. This relieves shoppers of the burden of having to formulate the perfect search query, says CTO Mike Fisher.
|
|
|
|
Dementia is one of the most significant global health concerns, with 75 million people projected to be living with it by 2030, according to Alzheimer’s Disease International. PHOTO: PHILIPPE LOPEZ/AGENCE FRANCE-PRESSE/GETTY IMAGES
|
|
|
AI may help identify patients with early-stage dementia. Researchers and companies are now testing whether machine learning can be used to help clinicians detect diseases such as Alzheimer’s sooner, Shirley S. Wang writes in The Wall Street Journal.
Case study. A study funded by the National Institute on Aging examined electronic records of more than 16,000 medical visits of 4,330 participants in a Kaiser Permanente Washington health system. Using a model that identified 31 factors associated with cognitive decline, researchers were able to flag more than 1,000 visits that resulted in a dementia diagnosis, including nearly 500 in which the patient’s cognitive changes previously had gone undetected in the health system.
Humans fail AI's potential in medicine. Research by Google Health in Thailand found potential for a deep-learning system designed to identify a diabetic eye disease. It didn't work out that way once the tech was transferred to the real world. "In some clinics, the lighting was not right, or internet speeds were too slow to upload the images," says the FT. "Some clinics did not use eye drops on patients, exacerbating problems with the images. The result was a large number of images that could not be analysed by the system." (FT)
AI listens to coughs, can detect Covid-19. Researchers at MIT say they have developed a system capable of recognizing the cough of someone with Covid-19. The system uses three neural networks: one to measure the vocal strength, a second to determine the emotional state of the individual and a third that can measure changes in respiratory performance. (Engadget)
Good night, algorithms. Singer and artist Grimes worked with Endel, an app that uses AI to build soundscapes to help users focus or relax, on an “AI Lullaby” aimed at helping people sleep. She says she was inspired by her five-month old son, X Æ A-XII Musk. Elon Musk is the father. (New York Times)
Soccer's all-header match. An AI-controlled camera in a Scottish soccer stadium programmed to follow the ball gave viewers extended views of a linesman's follicle-challenged head. (SBNation)
|
|
|
The U.S. Air Force recently bought dozens of Chinese-made drones to use for testing and training, fueling concerns about continued Defense Department use of technology that lawmakers consider a threat to national security. (WSJ)
European officials and privacy experts hope the U.S. presidential election opens the door to improving relations over data privacy that grew shaky in recent years after a series of court challenges to American intelligence programs. (WSJ)
|
|
|
|