Apple Buys ‘Dark Data’ Firm ‘For AI Training’

Lattice’s technology finds patterns in large amounts of unstructured data and could be used to train AI personal assistants

Apple confirmed it has acquired data mining company in a deal reportedly worth about $200 million (£155m).

The company’s technology is based on a Stanford project called DeepDive, a system for turning unstructured information into structured data that can be integrated into an existing structured database, and its co-founders are all academics who worked on the project.

Unstructured data

The technology could be used to help refine artificial intelligence software, a field that’s become big business in the form of personal assistants such as Apple’s Siri, Amazon’s Alexa, Google Assistant, Microsoft’s Cortana and Samsung’s Bixby.

Siri Logo
The deal took place about two weeks ago and will see about 20 engineers join Apple, according to a report by TechCrunch on Saturday.

Apple confirmed the acquisition. “Apple buys smaller technology companies from time to time and we generally do not discuss our purpose or plans,” the company said in a statement.

Lattice was founded in 2015 and raised at least $20m in venture capital funding before it first made itself publicly known last year.

Artificial intelligence

Its co-founders are Christopher Ré, a professor at Stanford and now Lattice chief scientist; Michael Cafarella, a professor at the University of Michigan, now its chief technical officer; Raphael Hoffman, now at Google; and Feng Niu, formerly a PhD student advised by Ré and now Lattice’s chief engineering officer.

In an online description of DeepDive Niu sit it could be useful for academics, investigators, people working in the legal field or other kinds of researchers.

A source told TechCrunch Lattice has been in talks with technology companies including Amazon and Samsung about using the tool to improve their artificially intelligent personal assistants.

Put your knowledge of artificial intelligence (AI) to the test. Try our quiz!