关键词: information decomposition natural language processing text style transfer

来  源:   DOI:10.3390/e25020322

Abstract:
A growing number of papers on style transfer for texts rely on information decomposition. The performance of the resulting systems is usually assessed empirically in terms of the output quality or requires laborious experiments. This paper suggests a straightforward information theoretical framework to assess the quality of information decomposition for latent representations in the context of style transfer. Experimenting with several state-of-the-art models, we demonstrate that such estimates could be used as a fast and straightforward health check for the models instead of more laborious empirical experiments.
摘要:
越来越多的关于文本风格转移的论文依赖于信息分解。通常根据输出质量凭经验评估所得系统的性能或需要费力的实验。本文提出了一个简单的信息理论框架,以评估在风格转移背景下潜在表示的信息分解质量。用几种最先进的模型进行实验,我们证明了这种估计可以用作模型的快速和直接的健康检查,而不是更费力的经验实验。
公众号