Peer-reviewed journal articles, conference papers, and preprints, grouped by year.
A Prabakaran, Priyesh Shukla
arXiv preprint arXiv:2602.15516
This preprint introduces a semantic-guided 3D Gaussian Splatting pipeline to remove transient objects and improve scene consistency in reconstructed environments.
D RV, L Kanawat, Priyesh Shukla
arXiv preprint arXiv:2602.14165
This work proposes a bidirectional 4 K analog signal-chain interface for superconducting qubit control and readout, enabling tighter coupling between cryogenic electronics and quantum processors.
N Nargund, Priyesh Shukla
arXiv preprint arXiv:2602.07374
TernaryLM introduces a memory-efficient language modeling approach with native 1-bit quantization and adaptive layer-wise scaling to retain quality while minimizing footprint.
Amit Ranjan Trivedi, Shamma Nasrin, Priyesh Shukla, et al.
IEEE 43rd VLSI Test Symposium (VTS), Tempe, AZ, USA
Compute-in-Memory (CiM) architectures, particularly those leveraging SRAM-based arrays, present significant opportunities for accelerating deep learning by mitigating data movement bottlenecks in traditional von Neumann systems. We propose MOSAIC, a novel CiM architecture designed around three foundational principles: co-designing deep learning operators for CiM, leveraging a memory-immersed digitization approach, and orchestrating inference over a network of compact CiM μArrays.
Priyesh Shukla, Amit Ranjan Trivedi, Siddarth Gottumukkula, M P Samartha
IEEE International Conference on Intelligent Computing and Systems at the Edge (ICEdge)
This work presents ProbCIMs, a compute-in-memory driven framework for robust and efficient probabilistic computing and AI workloads at the edge.
Nastaran Darabi, Priyesh Shukla, et al.
Design, Automation & Test in Europe Conference & Exhibition (DATE), Valencia, Spain
This paper addresses the challenging problem of energy-efficient and uncertainty-aware pose estimation in insect-scale drones. We introduce CIM-based acceleration of Bayesian filtering methods and variational inference of deep learning models through probabilistic processing.
Priyesh Shukla, Shamma Nasrin, et al.
IEEE Transactions on Circuits and Systems I: Regular Papers (TCAS I)
We propose MC-CIM, a compute-in-memory (CIM) framework for robust, yet low power, Bayesian edge intelligence. Using Monte Carlo Dropout approximation to Bayesian DNN, we discuss a novel CIM module that can perform in-memory probabilistic dropout in addition to in-memory weight-input scalar product.
Priyesh Shukla, Ankith Muralidhar, et al.
IEEE Transactions on Very Large Scale Integration (VLSI) Systems
We propose a novel compute-in-memory (CIM)-based ultralow-power framework for probabilistic localization of insect-scale drones. The proposed localization framework is ~25x energy-efficient than the traditional 8-bit digital GMM-based processor, paving the way for tiny autonomous drones.
Priyesh Shukla, Ahish Shylendra, et al.
IEEE International Symposium on Circuits and Systems (ISCAS), Seville, Spain
This work discusses the implementation of Markov Chain Monte Carlo (MCMC) sampling from an arbitrary Gaussian mixture model (GMM) within SRAM, enabling high performance Metropolis-Hastings algorithm-based MCMC sampling.