More AI for developers as Expect Labs releases the MindMeld API

Expect Labs, makers of the MindMeld app for dynamically suggesting content in response to the topics in a spoken conversation, is opening its artificial intelligence engine to the world via the new MindMeld API. It’s the latest example of just how powerful APIs are becoming and offers yet another glimpse into how intelligent we will expect applications to be in the years to come.

The key to the MindMeld API is its ability (well, the ability of the system behind it) to account for context. The API will index and make a knowledge graph from a website, database or content collection, but then it also collects contextual clues from an application’s users about where they are, what they’re doing or what they’re typing, for example. It’s that context that lets the API decide which search results to display or content to recommend, and when.

And although speech recognition was a big component of the MindMeld app, the API doesn’t require users to utilize a voice input. It could just as easily handle text search, location or other contextual inputs as the trigger for recommendations. Only about half the apps presently using the API rely on speech recognition, Tuttle said.

The API — which ranges from free to $1,999 per month — could potentially help someone build something similar to the Google Now assistant, but also a lot more. Maybe a news site wants to make better use of its video library by recommending videos to readers, or a call center wants to improve customer service by pointing operators to the right information to answer a customer’s question in real time. “As it turns out, developing this machine intelligence capability is core to a lot of businesses, not just Google’s,” Expect Labs co-founder and CEO Tim Tuttle said.

MindMeld listens to conversations and recommends relevant content.
The MindMeld app listens to conversations and recommends relevant content.

In fact, although the company’s flagship MindMeld app was a great way to show off what Expect Labs’ technology could do and to hone its capabilities, the API is the core to Expect Labs’ business going forward. “We’re trying to make it possible for this tech to get into every app, every device and every website,” Tuttle said.

Later, he added, “Five years from now, AI is going be the way that you get access to all the information that you want…it ends up being the gateway to the customer for just about everything you want.”

Expect Labs wants to be one of the biggest and best companies providing that gateway. Companies like Intel have invested in Expect Labs, Tuttle said, because they realize the shift that’s happening and they need to make sure their chips are designed to enable rather than hinder new capabilities.

Tim Tuttle

Tuttle, who’ll be speaking (alongside SwiftKey co-founder and CTO Ben Medlock) about the cutting edge of artificial intelligence at our Structure Data conference in March, said a convergence of factors is leading to an explosion in the AI space. A big one is the abundance and relative low cost of computing resources that can be accessed and delivered as cloud services. Another is the advent of data sets large enough to take advantage of techniques such as deep learning. (We’ll definitely touch upon this in other sessions, as well, including one dedicated to AI via APIs and another with D-Wave CEO Vern Brownell who’s promising quantum computing via API.)

“My whole career, I’ve been waiting for machines to finally demonstrate some human-like intelligence,” he said, and it looks as if that time is finally here.

Actually, Tuttle noted, although deep learning gets a lot of attention as the next big thing in AI and data processing, it’s not necessarily too advanced a technique and wasn’t always too useful. But as data sets blow up into billions of points and beyond, these systems start doing intelligent things that they didn’t do at smaller scales. “Ultimately, this has to be unsupervised [like many deep learning models are] because the data sets are so large,” he said.

However, the great thing — in theory — about new tools such as the MindMeld API is that developers without any prior expertise in such fields can build the capabilities into applications. Consumers who don’t care about the underlying technologies, but just want an app that works well, need never know what’s going on under the covers. “It’s gonna be baked into every app,” Tuttle said. “Every device that you use every day.”

Feature image courtesy of Shutterstock user agsandrew.