About Me
Welcome!
I’m a fifth-year Ph.D. Student at the Georgia Institute of Technology in Atlanta, Georgia, specializing in neuromorphic computing theory and algorithms. My work is situated within the dynamic and innovative Gigascale Reliable Energy-Efficient Nanosystem (GREEN) Lab, under the guidance of Prof. Saibal Mukhopadhyay.
Latest News
Here, I’ll share the most recent updates on my projects, publications, and any upcoming events or presentations. Check back often for the latest information!
- April 3, 2024: Happy to announce I received the Outstanding ECE Graduate Research Assistant Award
- April 1, 2024: Excited to announce I won the Best Presentation Award at Georgia Tech ECE Research Rally
- March 27, 2024: Happy to announce I won the Colonel Oscar P. Cleaver Award for my PhD proposal dissertation at the 2024 Roger P. Webb Awards Program in the School of Electrical and Computer Engineering. Huge thanks to my advisor Prof. Saibal Mukhopadhyay and committee members Prof. Suman Datta and Prof. Justin Romberg
- March 28, 2024: Excited to announce we have two papers accepted in L4DC, 2024. Thanks to my collaborators Hemant Kumawat, Beomseok Kang and Harshit Kumar.
- January 11, 2024: Delighted to receive the ECE STEER Fellowship, 2024
- January 16, 2024: Excited to share my paper, “Sparse Spiking Neural Network: Exploiting Heterogeneity in Timescales for Pruning Recurrent SNN”, is accepted in ICLR 2024 as a poster presentation.
Research Interests
My academic journey is driven by a critical examination of the limitations inherent in traditional deep neural networks (DNNs), such as significant energy demands, lack of adaptability, and poor temporal processing capabilities. In pursuit of more efficient, flexible, and temporally aware computational models, my work focuses on the exploration and advancement of spiking neural networks (SNNs).
Theoretical Foundations
My theoretical investigations are anchored in four key areas:
- Temporal Processing: Delving into the superior capabilities of SNNs in leveraging temporal information, offering new possibilities for processing sequences and time-dependent data.
- Event-based Processing: Exploring the unique efficiency and precision of SNNs in managing event-based data, essential for real-time and energy-efficient computations.
- Plasticity: Enhancing neural networks’ adaptability to dynamic environments and their resilience against noise and adversarial attacks, thereby creating systems that can adjust to new information while maintaining their integrity.
- Uncertainty: Developing methodologies to quantify and calibrate the uncertainty in SNN predictions, especially given their distinct temporal dynamics, aiming to bolster the reliability of SNN-based decision systems.
Practical Applications
Complementing my theoretical research, I am actively involved in applying these insights to revolutionize various fields:
- Large Scale Graph Optimization: Employing SNNs for complex optimization within large graphs, facilitating more efficient solutions in network analysis, logistics, and more.
- Ultra-fast Monte Carlo Simulations: Leveraging the speed and efficiency of SNNs to enhance Monte Carlo simulations and simulations of random walks, dramatically speeding up computational tasks across finance, physics, and engineering disciplines.
- Spike-based Event-driven Optimal Control: Innovating control systems powered by the dynamic and precise responses of SNNs to real-time events, setting new standards for speed and accuracy.
Through a balanced emphasis on theoretical exploration and practical application, my research endeavors to expand the horizons of neuromorphic computing and establish its foundations for addressing some of the most pressing challenges of our times.
Feel free to navigate through the website to learn more about my projects, publications, and academic pursuits. Your insights and inquiries are always welcome!