54.3 F
San Diego
Thursday, Mar 28, 2024
-Advertisement-

Ai-one’s ‘Artificial Brain’ Has a Real Eye for Data

Software writer ai-one Inc. doesn’t just promise code. The company promises to pull new perspectives and second opinions from seemingly inscrutable data.

“It tells you questions you didn’t know to ask,” said Olin Hyde, one of the company’s local executives.

The software company has a few employees in Zurich and a sister company in Berlin, but its heart is in a Bird Rock storefront — just a few steps from the Pacific Ocean. The place houses four people, including company President Thomas Marsh, as well as Hyde, who is vice president of business development.

The company’s product is Nathan, which it bills as an “artificial brain.” Hyde says the software is able to learn the task it has to complete by being exposed to just a few examples.

Nathan has uses in drug discovery, medical billing, social media and cybersecurity, Hyde said.

Software is Trainable

Imagine a person having to read through 1,000 legal contracts to find non-compete clauses. Such a clause might be worded in many ways, Hyde said, so the garden-variety keyword search might not be the best way to solve the problem.

“What our technology allows you to do is to give it a few examples, just like you would train a human,” Hyde said. Usually it takes less than 10 examples. Programming the software does not require “if-then” statements, he said.

In the legal example, the software can identify the documents containing a non-compete clause and those without, Hyde said. And if there’s a gray area, it will alert the user to that.

Electronic discovery is one of the ways the technology can make legal departments more productive. “We’re not replacing the lawyer; we’re getting rid of the monotonous work that the lawyer really doesn’t want to do,” Hyde said.

Ai-one doesn’t name its customers, but Hyde asserts they are out there.

One is a telecom carrier that uses the software to personalize the mobile phone experience. Another uses the software to read electronic medical records and apply the proper codes. A defense contractor uses the software to sift through the messages found on social media sites.

To show off the program’s capability, Hyde ran three articles about a Chinese Internet censorship incident through Nathan. The articles were from the People’s Daily of China, Al Jazeera and CNN. “Those are three really different worldviews,” Hyde said. The software was able to create distinct, fingerprintlike graphs of each article, illustrating keywords and associations. The graphs illustrate how different each news source was in covering the same event.

Things get really interesting, Hyde says, when you look for similarities or differences between those fingerprints.

In another example, Hyde pulls up a Chinese Wikipedia entry for the U.S. Navy. He notes the software was able to distinguish two meanings for “constitution.” There’s the legal document, and there’s the historic ship of the same name.

Complements Local Sectors

The business case for running the software, Hyde said, is gaining understanding from data and making better decisions.

The company’s business model is to license its technology to people who will then build applications for their respective fields. “We cannot possibly be experts in every single domain,” he said.

Some 180 investors worldwide raised $7.5 million for the company, Hyde said.

The business started in 2003 and got early customers in 2005, Hyde said. It introduced its software development kit in 2011 and has just come out with a cloud-based version.

In 2009, ai-one reincorporated as a Delaware C-corporation and came to San Diego.

Why San Diego? Marsh said UC San Diego has one of the top artificial intelligence schools in the country. He also said the company’s specialty is a good complement to San Diego’s analytics sector. On top of that, San Diego has the biotechnology, cybersecurity and federal business sectors.

Lifestyle and a good talent pool are other good reasons to be in San Diego, Hyde said.

The La Jolla foursome includes Kurt Manninen, a former IBM employee who now heads product development. Business co-founder and CEO Walter Diggelmann works in Zurich.

Hyde sees plenty of work ahead, since the amount of data in the world is increasing rapidly. He cites one recent research house’s estimate that it will grow fiftyfold between 2011 and 2015, and that humans will not be able to keep tabs on it all.

Hyde, incidentally, dismisses the phrase “big data.”

A better term might be “dark data,” he said, since no one knows what’s in it.

-Advertisement-

Featured Articles

Oberon Eyes Europe for Renewable DME

Leaders of Influence in Law 2024

-Advertisement-
-Advertisement-

Related Articles

-Advertisement-
-Advertisement-
-Advertisement-