Skip to main content

Allegro.ai is helping Hyundai mine the artificial intelligence gold rush

Hyundai Nexo

In November 2018, Hyundai announced an investment in an Israeli startup named Allegro.ai that specializes in deep learning-based computer vision. On the surface, it sounded like a standard business deal. Company A invests in company B. That’s great for both sides (hopefully, at least), but it has very few implications for the average consumer. This tie-up is a little bit different. While Hyundai owners won’t notice an immediate change, the partnership promises to allow the South Korean firm to bring more technology in its cars faster than if it had decided to do everything in-house.

Digital Trends talked to Nir Bar-Lev, Allegro.ai’s CEO, to find out more.

Recommended Videos

Hyundai is huge, it’s one of the largest car companies in the world, so why would it invest in a startup like Allegro.ai instead of developing the technology on its own? Building a car is hard, just ask any of the startups who have tried, struggled, and failed, but developing advanced software is also difficult and time-consuming.

Hyundai Ioniq autonomous concept

“There aren’t enough people with the right experience and knowledge to do deep learning at a high level. It’s kind of like trying to build a car in the stone age. The infrastructure that companies have at their disposal is somewhere along the lines of what existed 35-40 years ago in a traditional software industry,” Bar-Lev explained. He added that 99 percent of companies do not have the required expertise to work with deep learning. “If you think of the gold rush, everyone ultimately needed jeans and picks and shovels, otherwise they couldn’t mine the gold. It’s the same thing here.”

That’s where Allegro.ai comes in. While Hyundai will develop its own deep learning technology, its researchers will use Allegro.ai’s solutions to better understand how the puzzle pieces come together. “By making these tools commercially available, companies can access them, which means things are going to happen faster,” Bar-Lev predicted.

Teaching a car how to drive is a lot like teaching a teen how to drive in the sense that experience is key

The first (and most often cited) application of deep learning in the automotive world is powering an autonomous car. For it to work, a car needs to understand what it’s doing, what other cars are doing, and the type of environment it’s operating in. And, as Bar-Lev pointed out, driving a car in the United States is a completely different experience than driving in Abu Dhabi, or Guatemala City, or downtown Paris.

Teaching a car how to drive is a lot like teaching a teen how to drive in the sense that experience is key. For a 15-year old, the experience comes by spending hours behind the wheel next to an instructor. For a car, it requires feeding the software a mammoth amount of annotated data that teaches it what trees, trucks, and railroad crossings look like.

Allegro.ai doesn’t deal in data. The companies who want to build self-driving cars need to figure out how to gather it. It simply provides a platform that lets engineers annotate it and feed it to a car more efficiently and at scale. On a second but more lasting level, the same basic technology can be used to teach a car how to recognize who is in the car at any given time and what they are doing.

Stephen Edelstein/Digital Trends

“If a car is on-demand, it needs to somehow know what’s going on in the cabin. It needs to make sure that no one is littering the cabin, that no one is doing something they’re not supposed to be doing,” explained Bar-Lev. This type of technology is also used in semi-autonomous systems to tell whether the driver is looking at the road ahead, counting crows on power lines, or sleeping.

Finally, deep learning technology can also help automakers build better cars. Robots trained in quality control can identify even the tiniest scratches in the paint, misaligned body panels, or leaks before a car leaves the assembly line. Humans currently do this job. AI-enabled robots could replace them or complement them, depending on the company and the use case it makes for deep learning technology.

“A lot of folks don’t understand the need for deep learning, or why we’re talking about it. To the credit of [Allegro investors] Bosch and Samsung, they really saw where the market is going, and I think Hyundai joining us is a testament of the understanding across the industry,” Bar-Lev concluded.

Ronan Glon
Ronan Glon is an American automotive and tech journalist based in southern France. As a long-time contributor to Digital…
Researchers use artificial intelligence to develop powerful new antibiotic
MIT researchers used a machine-learning algorithm to identify a drug called halicin that kills many strains of bacteria. Halicin (top row) prevented the development of antibiotic resistance in E. coli, while ciprofloxacin (bottom row) did not.

Researchers at MIT have used artificial intelligence to develop a new antibiotic compound that can kill even some antibiotic-resistant strains of bacteria. They created a computer model of millions of chemical compounds and used a machine-learning algorithm to pick out those which could be effective antibiotics, then selected one particular compound for testing and found it to be effective against E. coli and other bacteria in mouse models.

Most new antibiotics developed today are variations on existing drugs, using the same mechanisms. The new antibiotic uses a different mechanism than these existing drugs, meaning it can treat infections that current drugs cannot.

Read more
MIT’s latest artificial intelligence can rewrite outdated Wikipedia pages
wikipedia asiacell iraq mwc2017 politics

Writers, editors, and researchers: Hold on to your red pens. Because MIT-powered A.I. may be coming for your jobs. 

A new “text-generating system” created by the brains behind Massachusetts Institute of Technology may be the beginning of the end for all human editing jobs. The system, announced in a press release Wednesday, is able to rummage through the millions of Wikipedia pages, sniff around for outdated data, and replace it with the most recent information available on the internet in a “human-like” style -- thus making the need for real, hot-blooded editors basically obsolete. 

Read more
Automated app helps ordinary citizens sue pesky robocall scammers

Nobody likes robocalls, the automated phone calls which are the bane of both the Federal Communications Commission and, well, every human with a phone. The average American citizen suffers through an estimated average of 18 robocalls every single month. In many cases, these robocalls try to scam the call recipient out of their hard-earned money, with losses adding up to billions of dollars each year. Could the answer to robocalls be more robots?

It may sound unlikely but Joshua Browder, the legal-tech whiz kid behind the DoNotPay parking ticket-fighting A.I., has created a new robo-tool with robocalls in its sights. Not only could it ultimately help stop you from receiving them; it could actually help you sue the callers for up to $3,000 per robocall they make. Heck, at those kinds of prices you might even grow to love the computerized auto-dial messages.

Read more