The chip has been used to demonstrate real-time learning with up to a 175x improvement in energy usage and it means that Chipzilla is just that much closer to having energy-efficient neuromorphic computing systems.

Intel Labs working with boffins from the Italian Institute of Technology and the Technical University of Munich, has apparently come up with a new way of working with neural network-based object learning.

It targets future applications like robotic assistants that interact with unconstrained environments, including in logistics, healthcare or elderly care. It uses neuromorphic using interactive online object learning methods to enable robots to learn new objects after deployment.

The boffins ran a spiking neural network architecture on Loihi that localised learning to a single layer of plastic synapses and accounted for different object views by recruiting new neurons on demand. This enabled the learning process to unfold autonomously while interacting with the user.

Intel research lead Yulia Sandamirskaya said that when a human learns a new object, they take a look, turn it around, ask what it is, and then they’re able to recognise it again in all kinds of settings and conditions instantaneously.

"Our goal is to apply similar capabilities to future robots that work in interactive settings, enabling them to adapt to the unforeseen and work more naturally alongside humans. Our results with Loihi reinforce the value of neuromorphic computing for the future of robotics," she said.