Condition-Wise Sinkhorn Drifting for One-Shot Learned Channel Simulation

2026-06-16

AI Summary

This work addresses the high computational cost of conventional diffusion models in learnable communication systems, which stems from repeated calls to channel simulators. To overcome this limitation, the authors propose a one-shot generative channel surrogate that efficiently models the conditional output distribution corresponding to a given transmitted symbol while preserving the input symbols unchanged. The approach innovatively integrates conditional output distribution modeling with Sinkhorn-based optimal transport, introducing a conditional Sinkhorn objective, finite-sample barycentric velocity training, and decoupled particle regression. Experimental results across AWGN, Rayleigh fading, SSPA nonlinear, and TDL channels demonstrate that the proposed method outperforms existing one-step drift variants in both conditional fidelity and symbol-level coding performance, making it particularly suitable for scenarios requiring frequent channel invocations.

Abstract

Learned communication systems may evaluate stochastic channel surrogates millions of times inside differentiable training loops, making diffusion-style reverse sampling expensive. This paper proposes condition-wise Sinkhorn drifting, a one-shot channel surrogate that preserves the transmitted symbol and transports only the conditional output laws \(p(y\mid x)\). We formulate a conditional Sinkhorn objective over repeated outputs at the same transmitted symbol and train the generator with finite-sample barycentric velocities followed by detached particle regression. Experiments on additive white Gaussian noise (AWGN), Rayleigh fading, solid-state power amplifier (SSPA) nonlinearity, and a compact tapped-delay-line (TDL) channel compare direct drifting, joint Sinkhorn drifting, condition-wise Sinkhorn drifting, conditional denoising diffusion probabilistic modeling (DDPM), denoising diffusion implicit modeling (DDIM), and Wasserstein generative adversarial network (WGAN) references. Within the evaluated one-shot drifting-family variants, condition-wise Sinkhorn is strongest under conditional diagnostics and symbolic-coding checks, while diffusion remains strongest on the hardest downstream symbol-error-rate (SER) curves. The resulting operating point is a condition-preserving one-shot simulator for settings where repeated channel calls make diffusion-style sampling too costly.

Problem

• learned communication systems
• channel simulation
• diffusion sampling
• one-shot generation
• conditional modeling

Innovation

• condition-wise Sinkhorn drifting
• one-shot channel simulation
• conditional optimal transport
• learned communication systems
• differentiable channel surrogate