arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.12857cs.ITmath.IT
收藏

基于A* 编码的快速相对熵编码

Fast Relative Entropy Coding with A* coding

Gergely Flamich, Stratis Markou, José Miguel Hernández-Lobato

摘要:

相对熵编码(REC)算法使用建议分布$P$对来自目标分布$Q$的样本进行编码,使得期望的码长是$\mathcal{O}(D_{KL}[Q \,||\, P])$。REC可以与现有的学习压缩模型无缝集成,因为与熵编码不同,它不假设离散的$Q$或$P$,并且不需要量化。然而,一般的REC算法需要一个棘手的$Ω(e^{D_{KL}[Q \,||\, P]})$运行时。介绍了基于A* 采样的两种REC算法AS* 和AD* 编码。我们证明了,连续分布在$\mathbb{R}$,如果密度比是单峰,AS* 有$\mathcal{O}(D_{\infty}[Q \,||\, P])$预期的运行时间,其中$D_{\infty}[Q \,||\, P]$是雷尼$\infty$ -分歧。我们提供的实验证据表明,AD* 也有$\mathcal{O}(D_{\infty}[Q \,||\, P])$预期的运行时间。我们证明AS* 和AD* 达到预期的码长$\mathcal{O}(D_{KL}[Q \,||\, P])$。此外,我们介绍了DAD*,一个近似算法的基础上AD*,它保留了其有利的运行时间,并有类似的替代方法的偏见。针对VAE,我们提出了IsoKL VAE(IKVAE),它可以与DAD* 一起使用,以进一步提高压缩效率。我们评估了A* 编码与(IK)VAE在MNIST上,它可以losslightly压缩图像接近理论上的最佳限制。

英文摘要:

Relative entropy coding (REC) algorithms encode a sample from a target distribution $Q$ using a proposal distribution $P$, such that the expected codelength is $\mathcal{O}(D_{KL}[Q \,||\, P])$. REC can be seamlessly integrated with existing learned compression models since, unlike entropy coding, it does not assume discrete $Q$ or $P$, and does not require quantisation. However, general REC algorithms require an intractable $Ω(e^{D_{KL}[Q \,||\, P]})$ runtime. We introduce AS* and AD* coding, two REC algorithms based on A* sampling. We prove that, for continuous distributions over $\mathbb{R}$, if the density ratio is unimodal, AS* has $\mathcal{O}(D_{\infty}[Q \,||\, P])$ expected runtime, where $D_{\infty}[Q \,||\, P]$ is the Rényi $\infty$-divergence. We provide experimental evidence that AD* also has $\mathcal{O}(D_{\infty}[Q \,||\, P])$ expected runtime. We prove that AS* and AD* achieve an expected codelength of $\mathcal{O}(D_{KL}[Q \,||\, P])$. Further, we introduce DAD*, an approximate algorithm based on AD* which retains its favourable runtime and has bias similar to that of alternative methods. Focusing on VAEs, we propose the IsoKL VAE (IKVAE), which can be used with DAD* to further improve compression efficiency. We evaluate A* coding with (IK)VAEs on MNIST, showing that it can losslessly compress images near the theoretically optimal limit.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑