arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.12258cs.CLcs.AI
收藏

推进文本可读性评估:Transformer 遇见手工构建的语言特征

Pushing on Text Readability Assessment: A Transformer Meets Handcrafted Linguistic Features

Univ. of Pennsylvania(宾夕法尼亚大学);Univ. of Wisconsin-Madison(威斯康星大学麦迪逊分校);LXPER AI(LXPER人工智能公司)

Bruce W. Lee, Yoo Sung Jang, Jason Hyung-Jong Lee

AI总结:

本文提出结合 RoBERTa 与 Random Forest 及 255 个手工语言特征的混合可读性评估模型,并引入三个高级语义新特征,在常用数据集上取得 SOTA,最高准确率达 99%。

摘要:

我们报告了可读性评估中的两项重要改进:1. 高级语义层面的三个新颖特征;2. 及时的证据表明,传统机器学习模型(例如使用手工特征的 Random Forest)能够与 transformer(例如 RoBERTa)结合,以增强模型性能。首先,我们探索合适的 transformer 和传统机器学习模型。随后,我们使用自主开发的提取软件提取 255 个手工构建的语言特征。最后,我们将这些特征组合起来,创建若干混合模型,在可读性评估的常用数据集上取得最先进(SOTA)准确率。手工特征的使用有助于提升模型在较小数据集上的性能。值得注意的是,我们的 RoBERTA-RF-T1 混合模型达到接近完美的 99% 分类准确率,较此前 SOTA 提高 20.3%。

英文摘要:

We report two essential improvements in readability assessment: 1. three novel features in advanced semantics and 2. the timely evidence that traditional ML models (e.g. Random Forest, using handcrafted features) can combine with transformers (e.g. RoBERTa) to augment model performance. First, we explore suitable transformers and traditional ML models. Then, we extract 255 handcrafted linguistic features using self-developed extraction software. Finally, we assemble those to create several hybrid models, achieving state-of-the-art (SOTA) accuracy on popular datasets in readability assessment. The use of handcrafted features help model performance on smaller datasets. Notably, our RoBERTA-RF-T1 hybrid achieves the near-perfect classification accuracy of 99%, a 20.3% increase from the previous SOTA.

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。

↑