arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2005.13455cs.CL
收藏

基于PRPN的无监督语法分析自训练

Self-Training for Unsupervised Parsing with PRPN

New York University(纽约大学);University of Colorado Boulder(科罗拉多大学博尔德分校)

Anhad Mohananey, Katharina Kann, Samuel R. Bowman

摘要:

神经无监督解析(UP)模型学习在不访问语法注释的情况下进行解析,同时针对语言建模等另一项任务进行优化。在这项工作中,我们提出了神经UP模型的自我训练:我们利用模型副本预测的聚合注释作为未来副本的监督。为了能够在训练期间使用我们的模型的预测,我们扩展了最近的神经UP架构,PRPN(Shen et al.,2018a),这样它就可以以半监督的方式进行训练。然后,我们将模型预测的解析示例添加到未标记的UP训练数据中。我们的自训练模型比PRPN高出8.1%F1,比以前的最先进模型高出1.6%F1。此外,我们表明,我们的架构也可以在超低资源设置的半监督解析有帮助。

英文摘要:

Neural unsupervised parsing (UP) models learn to parse without access to syntactic annotations, while being optimized for another task like language modeling. In this work, we propose self-training for neural UP models: we leverage aggregated annotations predicted by copies of our model as supervision for future copies. To be able to use our model's predictions during training, we extend a recent neural UP architecture, the PRPN (Shen et al., 2018a) such that it can be trained in a semi-supervised fashion. We then add examples with parses predicted by our model to our unlabeled UP training data. Our self-trained model outperforms the PRPN by 8.1% F1 and the previous state of the art by 1.6% F1. In addition, we show that our architecture can also be helpful for semi-supervised parsing in ultra-low-resource settings.

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。

↑