Scalable Discrete-to-Continuous Channel Simulation for Compression and Privacy
Abstract
Channel simulation has recently emerged as a useful component in machine learning systems where samples from a prescribed probability distribution are to be compressed. Yet, general channel simulation algorithms often suffer from high computational costs, random stopping times or, in the worst case, can require generating an infinite number of shared random samples. We introduce a scheme for both exact and approximate simulation of discrete-to-continuous channels which conversely uses a fixed number of random samples, and therefore has a runtime independent of the channel and the input. Unlike existing channel simulation schemes which generate a sequence of independent samples from a proposal distribution, our approach generates one sample, or alternatively a fixed number of samples, from each potential target distribution. We then apply a latent permutation to the samples before performing sample selection using an exponential race. Our scheme provides a flexible tradeoff between the number of generated samples and the compression rate. Using polar and multilevel coding, we scale our approach to handle long blocklengths in O(n log n) time in order to benefit from reduced per-symbol overhead. We conclude by demonstrating applications to variable-rate compression with stochastic VQ-VAEs and communication-efficient differentially private distributed mean estimation via exact simulation of the Gaussian mechanism.
Related
- Scalable Distributed Stochastic Optimization via Bidirectional Compression: Beyond Pessimistic Limits
- Exact Common Information and Exact Channel Synthesis for Correlated Discrete Sources
- Dense Structural Compression of Transformers via Gauge-Correct Channel Removal
- CONTINUOUS VERSUS DISCRETE STRUCTURES -DISCRETE EMBEDDING FORMALISMS AND CONTINUALIZATION METHODS
- Fast 3D diffusion for scalable discrete granular media synthesis
BibTeX
@misc{rowan2026scalable,
title = {{Scalable Discrete-to-Continuous Channel Simulation for Compression and Privacy}},
author = {Joseph Rowan and Buu Phan and Ashish J. Khisti},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.12067},
eprint = {2609.12067},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.12067},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Scalable Discrete-to-Continuous Channel Simulation for Compression and Privacy
AU - Joseph Rowan
AU - Buu Phan
AU - Ashish J. Khisti
PY - 2026
DA - 2026/09/10
PB - arXiv
DO - 10.48550/arxiv.2609.12067
UR - https://arxiv.org/abs/2609.12067
AB - Channel simulation has recently emerged as a useful component in machine learning systems where samples from a prescribed probability distribution are to be compressed. Yet, general channel simulation algorithms often suffer from high computational costs, random stopping times or, in the worst case, can require generating an infinite number of shared random samples. We introduce a scheme for both exact and approximate simulation of discrete-to-continuous channels which conversely uses a fixed number of random samples, and therefore has a runtime independent of the channel and the input. Unlike existing channel simulation schemes which generate a sequence of independent samples from a proposal distribution, our approach generates one sample, or alternatively a fixed number of samples, from each potential target distribution. We then apply a latent permutation to the samples before performing sample selection using an exponential race. Our scheme provides a flexible tradeoff between the number of generated samples and the compression rate. Using polar and multilevel coding, we scale our approach to handle long blocklengths in O(n log n) time in order to benefit from reduced per-symbol overhead. We conclude by demonstrating applications to variable-rate compression with stochastic VQ-VAEs and communication-efficient differentially private distributed mean estimation via exact simulation of the Gaussian mechanism.
ER -