Recent works have shown that the task of wireless transmission of images can
be learned with the use of machine learning techniques. Very promising results
in end-to-end image quality, superior to popular digital schemes that utilize
source and channel coding separation, have been demonstrated through the
training of an autoencoder, with a non-trainable channel layer in the middle.
However, these methods assume that any complex value can be transmitted over
the channel, which can prevent the application of the algorithm in scenarios
where the hardware or protocol can only admit certain sets of channel inputs,
such as the use of a digital constellation. Herein, we propose DeepJSCC-Q, an
end-to-end optimized joint source-channel coding scheme for wireless image
transmission, which is able to operate with a fixed channel input alphabet. We
show that DeepJSCC-Q can achieve similar performance to models that use
continuous-valued channel input. Importantly, it preserves the graceful
degradation of image quality observed in prior work when channel conditions
worsen, making DeepJSCC-Q much more attractive for deployment in practical
systems.

提出一种基于源信号与信道联合编码、优化的机器学习方法 DeepJSCC-Q，用于无线图像传输，能够在信道输入固定的情况下实现与传统连续值信道输入方法相似的性能表现，并保持图像质量的优雅退化特性，具有更多实际系统的部署价值。

DeepJSCC-Q：通道输入受限的深度联合源信道编码

DeepJSCC-Q: Channel Input Constrained Deep Joint Source-Channel Coding

This paper studies the haplotype assembly problem from an information
theoretic perspective. A haplotype is a sequence of nucleotide bases on a
chromosome, often conveniently represented by a binary string, that differ from
the bases in the corresponding positions on the other chromosome in a
homologous pair. Information about the order of bases in a genome is readily
inferred using short reads provided by high-throughput DNA sequencing
technologies. In this paper, the recovery of the target pair of haplotype
sequences using short reads is rephrased as a joint source-channel coding
problem. Two messages, representing haplotypes and chromosome memberships of
reads, are encoded and transmitted over a channel with erasures and errors,
where the channel model reflects salient features of high-throughput
sequencing. The focus of this paper is on the required number of reads for
reliable haplotype reconstruction, and both the necessary and sufficient
conditions are presented with order-wise optimal bounds.

该研究从信息论的角度探讨了单倍型组装问题，发现 DNA 测序技术提供的短读序列可以用于恢复目标单倍型序列。研究将这一问题重新表述为联合源通道编码问题，并给出了最佳边界的充分必要条件以及所需短读的数量，重点是可靠的单倍型重建。