Presented my research on local learning rules at WACV 2025!

I had an amazing time presenting my work at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025! I showcased our research on LLS: Local Learning Rule for Deep Neural Networks Inspired by Neural Activity Synchronization.

This work introduces a novel bio-inspired local learning rule that enables training deep networks without relying on backpropagation. By leveraging synchronization mechanisms observed in biological neurons, LLS and its variants (LLS-M and LLS-MxM) achieve competitive accuracy with up to 300× fewer MAC operations and significantly lower memory requirements—making them well-suited for on-device and edge AI applications.

It was a great opportunity to connect with researchers pushing the boundaries of efficient learning and edge intelligence.

A heartfelt thank you to my co-authors, Arani Roy and Prof. Kaushik Roy, to the Nanoelectronics Research Laboratory (NRL) at Purdue University, and Center for the Co-Design of Cognitive Systems (CoCoSys) for their incredible support throughout this project.

Check out the preprint on arXiv: https://arxiv.org/abs/2405.15868