9 papers
AgBench: Agentic AI Benchmarks for Personal AI Devices
Yizhou Han, Di Wu, Dhananjay Saikumar +1
Agentic AI systems increasingly rely on cloud-hosted large language models for planning, tool use, and iterative execution, raising concerns about API cost and data exposure. Advan…
Data-Free Pruning of Self-Attention Layers in LLMs
Dhananjay Saikumar, Blesson Varghese
Many self-attention sublayers in large language models (LLMs) can be removed with little to no loss. We attribute this to the Attention Suppression Hypothesis: during pre-training,…
Signal Collapse in One-Shot Pruning: When Sparse Models Fail to Distinguish Neural Representations
Dhananjay Saikumar, Blesson Varghese
Neural network pruning is essential for reducing model complexity to enable deployment on resource constrained hardware. While performance loss of pruned networks is often attribut…
DRIVE: Dual Gradient-Based Rapid Iterative Pruning
Dhananjay Saikumar, Blesson Varghese
Modern deep neural networks (DNNs) consist of millions of parameters, necessitating high-performance computing during training and inference. Pruning is one solution that significa…
Exploring the Frontiers: Challenges and Theories Beyond the Standard Model
Dhananjay Saikumar
Quantum Field Theory (QFT) forms the bedrock of the Standard Model (SM) of particle physics, a powerful framework that delineates the fundamental constituents and interactions of t…
NeuroFlux: Memory-Efficient CNN Training Using Adaptive Local Learning
Dhananjay Saikumar, Blesson Varghese
Efficient on-device Convolutional Neural Network (CNN) training in resource-constrained mobile and edge environments is an open challenge. Backpropagation is the standard approach…
Quantum Tunneling Insights into the Atomic Landscapes of Graphite, Gold, and Silicon
Dhananjay Saikumar
Scanning Tunneling Microscopy (STM) is a powerful technique that utilizes quantum tunneling to visualize atomic surfaces with high precision. This study presents detailed topograph…
Future of computing at the Large Hadron Collider
Dhananjay Saikumar
High energy physics (HEP) experiments at the LHC generate data at a rate of Terabits per second. This data rate is expected to exponentially increase as experimen…
Parallel N-body simulations of planetary systems: a direct approach
Dhananjay Saikumar
Direct gravitational simulations of n-body systems have a time complexity O(n^2), which gets computationally expensive as the number of bodies increases. Distributing this workload…