IIT Guwahati develops energy-efficient, brain-inspired AI model

Researchers at the Indian Institute of Technology (IIT) Guwahati have developed a brain-inspired Artificial Intelligence model designed to process long sequences of data with significantly lower energy consumption than conventional AI architectures. The model, titled ‘Spiking Heterogeneous Harmonic Resonate-and-Fire State Space Model’ (SH2RFSSM), was developed by the SustainAI Lab within the Mehta Family School of Data Science and Artificial Intelligence.

Unlike standard neural networks that continuously process information, the SH2RFSSM utilizes spiking neural networks that mimic biological neurons by activating only when meaningful events occur. This approach enables sparse, energy-efficient computation, which is particularly beneficial for battery-powered and resource-constrained devices.

The research, which was presented at the International Conference on Machine Learning (ICML) 2026 in Seoul, South Korea, has potential applications in several sectors, including wearable health monitoring, Internet of Things (IoT) sensors, smart manufacturing, environmental monitoring, autonomous systems, and long-term forecasting. By reducing computational load, the technology could enable direct on-device AI processing without heavy reliance on cloud computing.