New Delhi: A team of researchers from the Indian Institutes of Technology (IIT) Delhi and Bombay built a new neuromorphic hardware using magnetic materials. While interactive voice assistants like Alexa and Siri have made it very easy to obtain a response to a query or get some tasks done online, the procedure in the background is quite intricate.
It involves delivering the speech data to the cloud over the internet, processing the data in the cloud, and then providing the response to the device. But due to slow internet speed, power outages, and other factors, speech data is frequently lost. This may change with the advent of neuromorphic computing, where local hardware with built-in memory executes all operations, hence preventing data loss while simultaneously reducing power consumption.
Also read: IIT Madras ties up with Denmark for research on next-gen fuels, energy systems
The new neuromorphic hardware developed using magnetic materials is capable of storing data even when the power is off. It functions similar to a synapse in the neurological system, said the research team led by Prof. Pranaba Kishor Muduli at IIT Delhi and Prof. Debanjan Bhowmik at IIT Bombay.
This is the first experimental demonstration of a neuromorphic device using magnetic materials from India, according to the researchers. They detailed their device in the ACS Applied Electronic Materials journal. "We have high hopes that the study contributes significantly to both the India Semiconductor mission as well as Make-in-India mission of the government of India," said Prof Pranaba Kishor Muduli, Department of Physics, IIT Delhi, in a statement.
The device consists of an ultrathin layer of cobalt. Thickness of the layer used is in the nano-metre range, which is 80,000-100,000 times smaller than human hair. The fabrication needs enormous precision; hence it is carried out in an ultra-high vacuum chamber containing nearly no air molecules, the team said. (IANS)
(This story has not been edited by ETV Bharat and is auto-generated from a syndicated feed.)