Nervana Systems, another player in the suddenly hot “deep learning” space, has closed its second round of capital in the last four months.
The San Diego startup said it raised $3.3 million in Series A funding led by DFJ, which comes on top of a $600,000 seed round in April.
DFJ’s Steve Jurvetson will take a seat on the company’s board as part of the latest investment. Allen & Co., AME Cloud Ventures and Fuel Capital also participated.
Deep learning is a form of artificial intelligence that researchers have credited with recent leaps in areas like speech recognition and image search. That has sparked growing interest in Silicon Valley, with Google, Facebook and Twitter making notable acquisitions or hires in recent months and various prominent players betting their own money on the space.
As Re/code explained in an earlier piece:
Deep learning is a form of machine learning in which researchers attempt to train computer algorithms to spot meaningful patterns by showing them lots of data, rather than trying to program in every rule about the world. Taking inspiration from the way neurons work in the human brain, deep learning uses layers of algorithms that successively recognize increasingly complex features — going from, say, edges to circles to an eye in an image.
Notably, these techniques have allowed researchers to train algorithms using unstructured data, where features haven’t been laboriously labeled by human beings ahead of time. It’s not a new concept, but recent refinements have resulted in significant advances over traditional AI approaches.
Nervana is aiming to distinguish itself in the nascent field by focusing on building hardware optimized for deep learning software — and vice versa.
“Traditionally, people write software algorithms that work well on the hardware that exists,” Nervana CEO Naveen Rao said in an interview. “We have the ability to co-develop and co-optimize them both. We’re innovating on the hardware and software sides.”
“We’re going to take things to the next level by applying technologies with an order of magnitude more performance and scale than we can do today.”
This article originally appeared on Recode.net.