A graph showing how the technology combines the core software needed to drive AI with image capture hardware in a single electronic device. Photo credit: RMIT University
Prototype Tech is scaling down AI to provide brain-like functions in one powerful device.
Researchers have developed artificial intelligence technology that combines imaging, processing, machine learning, and storage into an electronic chip powered by light.
The prototype downsizes artificial intelligence technology by mimicking the way the human brain processes visual information. The nanoscale advancement combines the core software needed to control artificial intelligence with image capture hardware in a single electronic device.
With further development, the light-driven prototype could enable more intelligent and smaller autonomous technologies such as drones and robotics as well as intelligent wearables and bionic implants such as artificial retinas.
The study by an international team of Australian, American and Chinese researchers led by RMIT University was published in the journal Advanced materials.
Lead researcher Associate Professor Sumeet Walia of the RMIT said the prototype offers brain-like functions in a powerful device.
The light-powered AI chip – prototype technology that combines imaging, processing, machine learning and storage. Photo credit: RMIT University
“Our new technology radically increases efficiency and accuracy by bringing multiple components and functions together on a single platform, ”said Walia.
“It brings us closer to an all-in-one AI device that is inspired by nature’s greatest computing innovation – the human brain.
“Our goal is to reproduce a core feature of the brain’s learning by imprinting sight as memory. The prototype we developed is a major step forward in the direction of neurorobotics, better technologies for human-machine interaction, and scalable bionic systems. ”
Complete package: advanced AI
Typically, artificial intelligence depends heavily on software and external data processing. The new prototype aims to integrate electronic hardware and intelligence together for quick decisions on site.
“Imagine a dash cam in a car that is built into our neuro-inspired hardware. This means it can see lights, signs and objects and make instant decisions without having to connect to the Internet, ”said Walia, co-head of the Functional Materials and Microsystems Research Group at RMIT.
“By bringing everything together in one chip, we can achieve an unprecedented level of efficiency and speed in autonomous and AI-driven decisions.”
Associate Professor Sumeet Walia and Dr. Taimur Ahmed, RMIT University. Photo credit: RMIT University
The technology builds on an earlier prototype chip from the RMIT team that used light to create and modify memories.
Thanks to the new built-in functions, the chip can now capture images and automatically improve them, classify numbers and be trained to recognize patterns and images with an accuracy of over 90%.
The device is also easily compatible with existing electronics and silicon technologies, allowing for easy future integration.
See the light: how the technology works
The prototype is inspired by optogenetics, an emerging tool in biotechnology that allows scientists to immerse themselves with great precision in the body’s electrical system and use light to manipulate neurons.
The AI chip is based on an ultra-thin material – black phosphorus – that changes electrical resistance in response to different wavelengths of light. The various functions such as imaging or storage are achieved by directing different colors of light onto the chip.
The study’s lead author, Dr. Taimur Ahmed from RMIT said light-based computing is faster, more accurate and uses far less energy than existing technologies.
“By packing so much core functionality into a compact, nano-scale device, we can broaden the horizons of machine learning and the integration of AI into smaller applications,” said Ahmed.
“Using our artificial retina chip, for example, would allow scientists to miniaturize this emerging technology and improve the accuracy of the bionic eye.
“Our prototype is a significant step forward towards the ultimate in electronics: a brain-on-a-chip that can learn from its environment just like us.”
This work was carried out in part at the Micro Nano Research Facility (MNRF) of the RMIT with support from the RMIT Microscopy and Microanalysis Research Facility (RMMF), the National Computational Infrastructure Australia (NCI) and the Multimodal Australian Sciences Imaging and Visualization Environment (MASSIVE). and Pawsey Supercomputing Facility.
Reference: “Fully Light-Controlled Memory and Neuromorphic Computation in Layered Black Phosphorus” by Taimur Ahmed, Muhammad Tahir, Mei Xian Low, Yanyun Ren, Sherif Abdulkader Tawfik, Edwin LH Mayes, Sruthi Kuriakose, Shahid Nawaz, Michelle JS Spencer, Hua Chen, Madhu Bhaskaran, Sharath Sriram and Sumeet Walia, November 17, 2020, Advanced materials.
DOI: 10.1002 / adma.202004207
