arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2009.06358cs.CLcs.LGstat.ML
收藏

基于句子连贯目标的语言生成研究

Improving Language Generation with Sentence Coherence Objective

Stanford University(斯坦福大学)

Ruixiao Sun, Jie Yang, Mehrdad Yousefzadeh

摘要:

有条件故事生成和上下文文本延续已经成为NLP领域越来越热门的话题。现有的模型往往倾向于输出逐渐偏离给定提示的文本段落。虽然生成的文本可能具有一定的复杂性和多样性,但很容易被人类识别为乱码。我们项目的目标是提高语言生成模型中句子之间的连贯性和一致性。我们的目标是解决这个问题,首先用GPT-2预训练模型训练一个句子对连贯性分类器,然后使用类似于REINFORCE算法的方法,用这个新的连贯性目标共同训练GPT-2语言模型。这种经过微调的语言模型能够生成以给定主题为条件的冗长段落,而不会偏离太多。该模型的简单性允许其适用于各种底层语言模型架构,因为它只修改预训练模型的最后一层。

英文摘要:

Conditional story generation and contextual text continuation have become increasingly popular topics in NLP community. Existing models are often prone to output paragraphs of texts that gradually diverge from the given prompt. Although the generated text may have a reasonable perplexity and diversity, it could easily be identified by human as gibberish. The goal of our project is to improve the coherence and consistency across sentences in a language-generation model. We aim to solve this issue by first training a sentence pair coherence classifier with GPT-2 pretrained model, and then co-train the GPT-2 language model with this new coherence objective using a method analogous to the REINFORCE algorithm. This fine-tuned language model is able to generate lengthy paragraph conditioned on a given topic without diverging too much. The simplicity of this model allows it to be applicable to a variety of underlying language model architecture since it only modifies the final layer of the pre-trained model.

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。

↑