by Researchers at the Indian Institute of Science (IISc) have made a breakthrough in the field of neuromorphic computing. It creates an analog computing platform inspired by the brain. Which can store and process information in an unprecedented 16,500 conduction states within the ANAM molecular film. You can store and process .
This neuromorphic platform may revolutionize AI functions such as training large-scale language models (LLMs) by bringing them to personal devices such as laptops and smartphones. These functions are currently limited to Resource-intensive data centers due to lack of energy-efficient hardware. As silicon electronics approaches saturation point. Brain-inspired accelerator design It can work with silicon chips to deliver faster, more efficient AI Sri Tosh Goswami, Associate Professor from the Center for Nano Science and Engineering (CeNSE), IISc, heads the research team. He explains, “Neuromorphic computing has faced multiple challenges for over a decade. With this discovery We are close to perfecting the system. which is a rare achievement.”
Researchers at the Indian Institute of Science (IISc) have developed a brain-inspired analog computing platform. It can store and process information on an astonishing 16,500 driver states within the molecular film. The advance, published today in Nature, is a major step over traditional digital computers in The storage and processing of data is limited to only two states. Such platforms may bring complex AI tasks, such as training large-scale language models (LLMs), to personal devices such as laptops and smartphones, bringing us closer to democratizing AI development. Current tools These developments are limited to resource-intensive data centers. This is due to the lack of energy efficient hardware. Silicon electronics is close to saturation. With an accelerator design inspired by the brain It can work with silicon chips to deliver faster, more efficient AI
Neuromorphic computing has its fair share of challenges that have remained unresolved for more than a decade, explains Sri Tosh Goswami, associate professor at the Center for Nanoscience Engineering (CeNSE), IISc, who led the research team. With this discovery We almost managed it perfectly. which is a rare achievement
The basic functions that underpin most AI algorithms are very basic. That is matrix multiplication. This is a concept taught in high school mathematics. However, in digital computers These calculations require a lot of energy. The platform developed by the IISc team significantly reduces the time and energy involved. This makes these calculations much faster and easier.
The molecular system at the heart of the platform was designed by Goswami, a visiting professor at CeNSE. As the molecules and ions move and move within the physical film, They create many different memory states. The basic function that forms the basis of most AI algorithms is matrix multiplication.] However, these calculations are energy intensive in digital computers. The platform developed by the IISc team significantly reduces the time and energy required. This makes these calculations much faster and more efficient.
The molecular mechanism at the heart of the platform was designed by Goswami using precisely timed voltage pulses. The team was able to track the movement of a much larger number of molecules. and map each electrical signature
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it
