IISc Develops Human Brain-Inspired Computing Platform

Source: This post on IISc Develops Human Brain-Inspired Computing Platform has been created based on the article “IISc proposes ‘human brain’ inspired computing platform to boost speeds, energy efficiency of AI tools”, published in Indian express on 14^th Sep 2024.

Why in news?

Researchers from the Indian Institute of Science (IISc), Bengaluru have created a revolutionary computing platform inspired by human brain function. The platform mimics brain-like processes and could significantly enhance AI tools’ speed and energy efficiency.

 About the ‘Brain on Chip’ Design

1. Developed by:  the Centre for Nano Science and Engineering (CeNSE) at IISc.

2. The system processes and stores data, resembling the functioning of a human brain. It offers 16,500 conductance states within a molecular film, far surpassing traditional digital systems which rely on binary (0 and 1) states.

3. Advantages Over Traditional Digital Platforms:  Existing digital platforms consume large amounts of energy and time, limiting speed. In contrast, this brain-like computing platform is much faster and more energy-efficient.

4. It allows AI-based tasks to be easily deployed on personal devices like smartphones, laptops, and desktops.

Features of brain-inspired neuromorphic computing platform

1. The brain-inspired neuromorphic computing platform’s molecular film enables free ionic movement, creating countless memory pathways and intermediary states, which are inaccessible in conventional digital platforms.

2. Digital platforms can typically only access two conductance states, but this new system offers precise control over molecular kinetics, mapping each movement to a distinct electrical signal.

3. The new system features a neuromorphic accelerator, which can store and process data in the same location, closely mimicking how the human brain operates. This is the first system of its kind, and its tiny molecular changes enable precise and efficient functioning.

Applications and Future Potential

The system could lead to breakthroughs in AI and neuromorphic computing, improving applications like Machine learning models, Edge training, Generative adversarial networks, Memory networks (long-term, short-term) and Transformers.

About Neuromorphic computing

1. It is an innovative field of computing that aims to mimic the architecture and functioning of the human brain.

2. Traditional computers operate using binary logic (0s and 1s) in sequential processing, while neuromorphic systems are designed to replicate the way neurons and synapses communicate in the brain.

3. These systems rely on analog signals and multiple conductance states, enabling them to process information in a manner that closely resembles biological neural networks.

4. Neuromorphic computing is particularly beneficial for AI tasks. By replicating neural networks and integrating processing and memory, these systems are able to perform computations with higher accuracy and efficiency.

UPSC Syllabus: Science and technology