This is the web version of Eye on A.I., Coins2Day’s weekly newsletter covering artificial intelligence and business. To get it delivered weekly to your in-box, sign up here.
What’s the new normal going to look like? It’s a vexing question for data scientists and engineers working on artificial intelligence systems.
So many of these systems are designed to learn from historical data. But, as I’ve noted in previous Coins2Day Eye on A.I. Newsletters, when the present stops looking like the past, these A.I. Systems can get into serious trouble.
What’s more, it’s unlikely that the future—when lockdowns ease and business resumes, but perhaps with some social distancing and travel restrictions still in place—will resemble today, either. So how can A.I. Software be trained to function correctly with all this changed data?
I put this question to Ahmer Inam, the chief A.I. Officer for Pactera Edge, a Redmond, Washington-based technology consulting and services firm that spun out from Chinese IT company Pactera Technology International in January. It helps businesses, including many Coins2Day 500 companies, implement A.I.
Inam answered that one way to make sure an A.I. System doesn’t run amok is to ensure a “human-in-the-loop” is always reviewing its recommendations and making final decisions. Today’s A.I. Systems function best when supporting human decisions, he says, not fully automating them.
Humans should also keep a close watch on the data that A.I. Software is being fed. That’s important even when there isn’t a pandemic. “Model drift,” in which data gradually changes over time, is “a common problem even with standard business data,” he says. “The dynamism of the business itself causes that.”
But another possible solution is to use a different kind of machine learning altogether. Instead of supervised learning, in which an algorithm learns from historical data, Inam says businesses could use reinforcement learning, in which an algorithm learns from experience, usually in a simulator.
This is the kind of A.I. That researchers have successfully used in the past five years to teach software to beat humans at games, such as Go and poker. But businesses have been slow to adopt these techniques for a number of reasons.
Building a reliable simulator can be expensive and time-consuming. The level of machine learning expertise needed to successfully design and train a reinforcement learning algorithm tends to be greater, and that talent is in short supply. The amount of computing power needed to train an algorithm using reinforcement learning can be massive and costly too.
Inam says the expense and time, however, are well worth it. Using a simulator, you can expose an algorithm to a wide range of potential scenarios and teach it how to handle them all. The result is a much more robust A.I., ready for whatever the world is going to throw at it.
Inam has built simulators to see how a coffee shop chain might be able to mitigate climate change’s affects on the reliability and pricing of coffee supplies. In another case, he simulated how a hurricane could impact an automotive retailer’s regional sales. More recently, he built a simulator to help a logistics company optimize its routing. The project took six months, he says, but it ultimately enabled the company to save millions in fuel and labor costs.
Inam thinks the coronavirus pandemic is likely to accelerate companies’ adoption of these more sophisticated A.I. Techniques. “It’s fast-tracking adoption of stronger A.I.,” he says. Add that to the list of the pandemic’s unanticipated effects.
***
While you’re here, we’d like your opinion on how we’re doing with the newsletter. Please spare the time to answer some quick questions for us. It will take no more than 2 minutes to complete.
We take your privacy seriously. The data we collect in this survey will be used for research purposes only.
Read on for the rest of this week’s A.I. News!
Jeremy Kahn
[email protected]
@jeremyakahn
This story has been updated to correct the relationship between Pactera Edge and Pactera Technology International.
A.I. IN THE NEWS
Self-driving startup settles lawsuit with Tesla. Zoox, a self-driving car startup in Foster City, California, has agreed to settle a lawsuit Tesla filed against it last year. The electric car giant had alleged four of its former employees had stolen confidential documents and trade secrets related to its logistics and warehouse operations and had taken them to Zoox. As part of the settlement, Zoox admitted to this allegation and agreed to pay Tesla an undisclosed amount as well as submit to audits to ensure Tesla's information wasn't being used, Reuters reported.
Covid-19 anxiety is getting worse, A.I. Reveals. Italian artificial intelligence company Expert Systems and Spanish political research firm Sociometrica have teamed up and applied natural language processing and sentiment analysis to 63,000 English-language social media posts. They found that, while positive sentiments have surged at moments, overall, fear and anxiety have continued to grow. You can see more of their analysis here.
Coursera uses A.I. To help universities offer online courses. The online learning company Coursera unveiled an A.I.-powered tool called "CourseMatch" to help universities use its service. The system ingests a university's entire course catalogue and then uses natural language processing to match each course to the five Coursera offerings that most closely align with that particular course description, according to a company blog post.
Amazon Alexa’s reading voice gets better. Amazon says it has enhanced the reading voice of popular digital assistant Alexa, creating a new speaking style. Amazon, according to a report in The Verge, has trained the digital assistant to read longer passages more like human speakers, adding "more natural pauses" between paragraphs or when the speaker switches during a dialogue. The function is currently only available in the U.S.
EYE ON A.I. TALENT
DBLX, an employee engagement startup based in Stoke on Trent, U.K., has hired Dan Macklin as chief technology officer. Macklin had previously been part of the U.K. Innovations team at Amazon Web Services.
Jvion, an Atlanta-based company implementing A.I. In clinical settings, has named Jay Deady as its chief executive officer. Deady had previously been CEO of A.I. Company Recondo.
Patriot One Technologies Inc., a Toronto company that uses artificial intelligence to help law enforcement agencies and the defense sector detect threats based on a variety of sensors and imagery data, has named Dietmar Wennemer as its chief operating officer. Wennemer had previously been chief product officer at Aeryon Labs, a drone company.
EYE ON A.I. RESEARCH
New specialist data sets for self-driving and satellite computer vision are out. Several companies are publicly releasing formerly proprietary datasets so that other A.I. Researchers can use them to build potentially useful applications.
First, Audi released a dataset of 400,000 images recorded on highways, country roads, and cities in the south of Germany along with accompanying sensor data. The images were recorded in a variety of different weather conditions. Audi also released other data recorded simultaneously with the images: three dimensional point clouds, bounding boxes, segmentation data and data on what vehicle did in response to those conditions. Other researchers could use this dataset, which Audi calls A2D2, to build new computer vision systems or enhance existing self-driving training datasets. Audi's self-driving rival Waymo previously made public a very large dataset.
Meanwhile, a nonprofit consortia called SpaceNet, whose backers include C.I.A.-affiliated venture capital firm In-Q-Tel, Capella Space, Maxar Technologies, German Aerospace Center, the Intel AI Lab and Amazon Web Services, has released a large dataset of images of the port of Rotterdam, Netherlands, and the surrounding area. The data, generated using satellite-based optical cameras and synthetic aperture radar, is extremely high-resolution, with precision down to less than a meter. SpaceNet has announced a competition around the data, asking researchers to "automatically extract building footprints with computer vision and artificial intelligence (AI) algorithms" using a combination of both the optical and radar data. The resulting A.I., SpaceNet says, could have uses in helping humanitarian groups and emergency services respond to disasters. Unspoken is that it could also be pretty useful to intelligence agencies.
FORTUNE ON A.I.
More surveillance and less privacy will be the new normal after the coronavirus pandemic—by David Meyer
Researchers working on ‘contact tracing’ say they welcome Apple and Google’s help—by David Z. Morris
The coronavirus crisis is fintech’s biggest test yet—and greatest opportunity to go mainstream—by Jen Wieczner
These robot-powered warehouses could save grocers—but first they need to survive the coronavirus pandemic—by Jeremy Kahn
Some of these stories require a subscription to access. Thank you for supporting our journalism.
BRAIN FOOD
In the opening of this newsletter, Pactera's Ahmer Inam told me that he thinks the coronavirus pandemic will accelerate the adoption of more sophisticated forms of A.I. Will it do the same for our embrace of robots?
Noted futurist Martin Ford thinks so. He tells the BBC that Covid-19 is changing our calculus about the advantages and disadvantages of human interaction and is going to "really open up new opportunities for automation." The fact that some form of social distancing will likely need to remain in place until a vaccine is widely available provides plenty of time and incentive for companies to speed up the deployment of robots.
In South Korea, robots are already being used to scan people's temperatures and distribute hand sanitizer, the BBC says, and a Danish company that makes robots that disinfect surfaces is doing brisk business. Businesses may now advance their deployment of robots that move items around warehouses and help prepare food. I recently talked to folks at IBM about some of their efforts to use A.I. To assist in situations such as helping elderly people to live independently.
But I am less sure about the pandemic is an unconditional win for robotics—the coronavirus may also have revealed many of the shortcomings of today's automation. Today's robotic systems have nowhere near the flexibility and resiliency of human workers: In a crisis, it is relatively easy to train a human factory worker to assemble parts for a ventilator instead of a carburetor, or redeploy a marketing executive to stack shelves in a supermarket. We are general-purpose machines. Robots—at least today's versions—are not. What do you think?
