Irish chip maker Movidius has created the world’s first deep learning USB stick that can add artificial intelligence (AI) to future products from self-driving cars to robots, and drones that will learn to think for themselves.
Entitled the Fathom Neural Compute Stick, the device will sell for less than $100 and will allow powerful neural networks to be moved out of the cloud and deployed on new products like robots and drones.
Think of it as AI on a stick.
It is the latest breakthrough for the Dublin company, which has been winning major multi-million dollar deals with Google and drone maker DJI.
‘With Fathom, every robot, big and small, can now have state-of-the-art vision capabilities’
– DR YANN LECUN, NEW YORK UNIVERSITY
“Any organisation can now add deep learning or machine intelligence to devices using the USB stick and create products that will be accessible to broader markets,” Movidius co-founder David Moloney told Siliconrepublic.com.
“We’ve already seen how the auto industry has been outflanked by Tesla and this is also starting to affect other industries. Knowledge-enabling products will make them more attractive to the consumer.
“The Fathom stick will bring machine intelligence to products to keep consumers engaged.”
New stick will democratise machine learning
Moloney said that the neural network capabilities and deep learning on the Fathom stick enable developers to create applications such as cameras that can track a face in a crowd, count the number of people in a crowd or track a car – you are limited only by your imagination.
When connected to a PC, the Fathom Neural Compute Stick behaves as a neural network profiling and evaluation tool. This means companies will be able to prototype faster and more efficiently and reduce the time-to-market for products that require AI.
‘What we are trying to do is democratise the technology and make it available to the next tier of inventors and product creators’
– DAVID MOLONEY, MOVIDIUS
Fathom allows developers to train neural networks and then deploy a low-power version to devices that contain a Myriad 2 processor.
“There are all kinds of things you can do with this,” Moloney said. “You could train the stick to be a kind of neural watchdog that watches for potential intrusions on your network, for example, to prevent data breaches.”
He said that it supports the major deep learning frameworks in use today such as Caffe and Tensorflow and that it is also compatible with Facebook’s own machine-to-machine programming language.
All robots big and small
In recent weeks, Movidius signed a multi-million dollar deal with Google that will see its chips bring human vision processing capabilities to future smartphones, as well as VR headsets.
Movidius has also been revealed as the brains behind the world’s first autonomous drone, the DJI Phantom 4.
Moloney said, for example, the USB device could be plugged into a robotic vacuum cleaner to enable the camera to interpret its environment and recognise furniture, pets and other obstacles, and learn from its mistakes.
“What we are trying to do is democratise the technology and make it available to the next tier of inventors and product creators.”
He said the key revolution is that makers of future machines and robots will be able to cost-effectively put self-learning and machine vision capabilities in new products.
“As a participant in the deep learning ecosystem, I have been hoping for a long time that something like Fathom would become available,” said founding director of New York University’s Data Science Centre, Dr Yann LeCun.
“The Fathom Neural Compute Stick is a compact, low-power convolutional net accelerator for embedded applications that is quite unique. As a tinkerer and builder of various robots and flying contraptions, I’ve been dreaming of getting my hands on something like the Fathom Neural Compute Stick for a long time. With Fathom, every robot, big and small, can now have state-of-the-art vision capabilities.”
Movidius, which was founded 10 years ago by Moloney and Sean Mitchell, recently raised €38m in a move that will enable it to generate 100 new jobs in Dublin.
The company now has offices in Silicon Valley, Ireland, Romania and China.
Robots image via Shutterstock