Using Gaze to Predict Text Readability

Using Gaze to Predict Text Readability
复制标题

使用注视来预测文本的可读性

DOI:
--
复制
发表时间:
2017
期刊:
BEA@EMNLP
影响因子:
--
通讯作者:
Anders Søgaard
Anders Søgaard
中科院分区:
--
文献类型:
--
作者:
Ana Valeria González;Anders Søgaard

文献摘要

被引文献

相似文献

我们发现,文本可读性预测显着提高从硬参数共享模型预测第一遍持续时间,总固定时间和回归持续时间。具体而言,我们诱导多任务多层感知器和逻辑回归模型的句子表示,捕捉各种汇总统计数据,从两个不同的文本可读性语料库的英语,以及邓迪眼动跟踪语料库。我们的方法比单任务学习和以前的系统有了显著的改进。此外,我们的改进在训练样本大小上是一致的,这使得我们的方法特别适用于小数据集。
We show that text readability prediction improves significantly from hard parameter sharing with models predicting first pass duration, total fixation duration and regression duration. Specifically, we induce multi-task Multilayer Perceptrons and Logistic Regression models over sentence representations that capture various aggregate statistics, from two different text readability corpora for English, as well as the Dundee eye-tracking corpus. Our approach leads to significant improvements over Single task learning and over previous systems. In addition, our improvements are consistent across train sample sizes, making our approach especially applicable to small datasets.