no code implementations • 22 Jan 2024 • Sadaf Salehkalaibar, Jun Chen, Ashish Khisti, Wei Yu
We derive the RDP function for vector Gaussian sources and propose a waterfilling type solution.
no code implementations • 23 Jan 2023 • Yangyi Liu, Stefano Rini, Sadaf Salehkalaibar, Jun Chen
This paper proposes ``\emph{${\bf M}$-magnitude weighted $L_{\bf 2}$ distortion + $\bf 2$ degrees of freedom''} (M22) algorithm, a rate-distortion inspired approach to gradient compression for federated training of deep neural networks (DNNs).
1 code implementation • 6 Feb 2022 • Sadaf Salehkalaibar, Stefano Rini
Under this assumption on the DNN gradient distribution, we propose a class of distortion measures to aid the design of quantizers for the compression of the model updates.
no code implementations • CVPR 2022 • Huan Liu, Zijun Wu, Liangyan Li, Sadaf Salehkalaibar, Jun Chen, Keyan Wang
Motivated by this observation, we propose a test-time training method which leverages a helper network to assist the dehazing model in better adapting to a domain of interest.