From drones, to surveillance cameras, to home appliances—devices of all kinds can become more powerful and useful by adopting artificial intelligence capabilities. However, today’s graphics processing units (GPUs) and cloud-based AI solutions can strain the capabilities and drain the batteries of these devices. Startup company Mythic believes it has the solution to this problem with a unique microchip that delivers capabilities like voice control and object recognition with minimal power consumption, according to Tech Crunch.
The Mythic platform performs hybrid digital/analog calculation inside flash memory arrays. This approach eliminates the need for expensive GPUs commonly used for local AI implementations. Mythic said its solution uses 50 times less energy than other solutions while delivering the processing power needed to run deep neural networks.
The solution also supports all common forms of neural networks, including convolutional, recurrent, and fully connected.
Mythic foresees its technology playing a role in a world where AI is added to every type of electronic device. For example, the company foresees autonomous drones capable of monitoring forest fires while simultaneously processing 4K video.
Moreover, Mythic envisions ubiquitous battery-powered monitors that can serve as electronic eyes and ears. These monitors also could work in concert, acting as security teams that could tell the difference between threats and everyday activity and react appropriately.
Other potential applications include a refrigerator that can serve as a meal planner, a fitness band that acts as a life coach and a smartphone that can perform the role of a medical specialist.
Stephan Cunningham is vice president, product management at Veritone. Working in concert with core internal teams including industry-specific general managers and engineering as well as directly with clients and prospects, he leads the disciplines and business processes which govern the Veritone Platform.