GUWAHATI: Researchers at IIT Guwahati have developed a brain-inspired artificial intelligence model designed to process long sequences of data while using less energy than conventional AI systems.
ALSO READ: Guwahati’s LNIPE Student’s Death Triggers Protest, Demands Probe Into Medical Care
The model, developed by researchers from the Mehta Family School of Data Science and Artificial Intelligence, was presented at the International Conference on Machine Learning (ICML) 2026 in Seoul, South Korea.
Named ‘Spiking Heterogeneous Harmonic Resonate-and-Fire State Space Model’ (SH²RFSSM), the system combines spiking neural networks with state space modelling to process sequential data more efficiently.
Researchers tested the model on 17 benchmark datasets covering areas including long-range sequence classification, human activity recognition, regression and long-term forecasting. They reported that it performed comparably to existing advanced sequence models while consuming substantially less estimated energy.
The model could have applications in wearable devices, IoT sensors, environmental monitoring, smart manufacturing and autonomous systems, particularly where AI needs to operate on battery-powered or resource-constrained devices.
The research was co-authored by Kartikay Agrawal, Vaishnavi Nagabhushana, Abhijeet Vikram, Vedant Sharma and Ayon Borthakur.