C Krishna Mohan

Minimizing Energy Costs in Deep Learning Model Training: The Gaussian Sampling Approach

2024 · arXiv (Cornell University)

Computing the loss gradient via backpropagation consumes considerable energy during deep learning (DL) model training. In this paper, we propose a novel approach to efficiently compute DL models' gradients to mitigate the substantial energy overhead …

Revamping Federated Learning Security from a Defender's Perspective: A Unified Defense with Homomorphic Encrypted Data Space

2024

Federated Learning (FL) facilitates clients to collaborate on training a shared machine learning model without exposing individual private data. Nonetheless, FL remains susceptible to utility and privacy attacks, notably evasion data poisoning and model inversion …