Reproduce “Rate-Distortion-Perception Tradeoff for Vector Gaussian Sources with Limited Common Randomness: Exact Characterization and Approximate Universality” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Wentao Wang, Shuo Shao, Ruibin Li, Jingjing Qian, Jun Chen, Bin Xia
The authors characterize joint communication and shared-randomness allocation for perceptual compression of vector Gaussian sources, balancing squared-error fidelity with a block Wasserstein-2 constraint. They derive an allocation rule and propose a fixed representation that allocates communication by reverse waterfilling, then randomness. Its distortion is at most 0.2897 times source variance above optimal, with an optimal-to-greedy ratio of at least 0.7517, within this model class.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.