arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 1904.09751cs.CL
收藏

神经文本退化的奇怪案例

The Curious Case of Neural Text Degeneration

University of Washington(华盛顿大学);Allen Institute for Artificial Intelligence(艾伦人工智能研究所);University of Cape Town(开普敦大学)

Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, Yejin Choi

摘要:

尽管深度神经语言模型取得了相当大的进步,但当这些模型作为文本生成器进行测试时,神经文本退化的谜团仍然存在。反直觉的经验观察是,即使使用似然性作为训练目标导致用于广泛的语言理解任务的高质量模型,使用似然性作为解码目标导致平淡且奇怪地重复的文本。在本文中,我们揭示了人类文本和机器文本之间令人惊讶的分布差异。此外,我们发现,解码策略本身可以显着影响机器文本的质量,即使从完全相同的神经语言模型生成。我们的发现激发了Nucleus Sampling,这是一种简单但有效的方法,可以从神经生成中获得最佳效果。通过从概率分布的动态核心中对文本进行采样,允许多样性,同时有效地截断分布的不太可靠的尾部,所产生的文本更好地展示了人类文本的质量,在不牺牲流畅性和连贯性的情况下增强了多样性。

英文摘要:

Despite considerable advancements with deep neural language models, the enigma of neural text degeneration persists when these models are tested as text generators. The counter-intuitive empirical observation is that even though the use of likelihood as training objective leads to high quality models for a broad range of language understanding tasks, using likelihood as a decoding objective leads to text that is bland and strangely repetitive. In this paper, we reveal surprising distributional differences between human text and machine text. In addition, we find that decoding strategies alone can dramatically effect the quality of machine text, even when generated from exactly the same neural language model. Our findings motivate Nucleus Sampling, a simple but effective method to draw the best out of neural generation. By sampling text from the dynamic nucleus of the probability distribution, which allows for diversity while effectively truncating the less reliable tail of the distribution, the resulting text better demonstrates the quality of human text, yielding enhanced diversity without sacrificing fluency and coherence.

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。

↑