Collaborative Research: Automatic Text-Simplification and Reading-Assistance to Support Self-Directed Learning by Deaf and Hard-of-Hearing Computing Workers

协作研究：自动文本简化和阅读辅助，支持聋哑和听力障碍计算工作者的自主学习

• 批准号: 2055699
• 负责人: Wei Xu
• 金额: $ 28.89万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-01 至 2023-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2055699&HistoricalAwards=false
• 关键词: Collaborative; Research; Automatic; Text; Simplification

While there is a shortage of computing and IT professionals in the U.S., there is underrepresentation of people who are Deaf and Hard of Hearing (DHH) in such careers. There is great diversity in the English reading skills among DHH Americans, who face challenges in the IT industry where regular upskilling is required throughout the working career. This interdisciplinary research will focus on studying automatic text-simplification technologies as reading assistance to help DHH individuals keep up with rapidly changing technologies through self-directed learning, outside of a formal classroom setting. The research team consists of experts in natural language processing (NLP), human-computer interaction (HCI), accessibility, and Deaf STEM education research. The resources and technologies developed for this project will be adaptable to other languages and text domains (e.g. medical information for lay readers), benefiting a wide range of populations (e.g. children, non-native speakers, people with reading disabilities).Prior text-simplification research for specific user groups focused on preliminary data collection or classical NLP methods, and little HCI research has explored design options for reading-assistance systems. This project will fill this critical gap in research and will investigate the learning and reading-assistance needs of upskilling DHH computing workers, design parameters influencing usability of reading-assistance tools, new training data and methods for text simplification tailored to specific reader groups and technical genres, automatic and human-based evaluation methods, and the impact of such tools on heutagogical learning. The methods will include interview and survey research with DHH computing workers, prototyping and testing of design variations, creation of parallel simplification corpora, readability annotation of lexicon and texts by DHH individuals, NLP research on domain adaptation and syntax-based neural machine translation for text simplification, and observation of real-world deployment of a prototype with DHH students and recent graduates from computing-related programs at the National Technical Institute for the Deaf (NTID).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

虽然美国缺乏计算和 IT 专业人员，但从事此类职业的聋人和听力障碍 (DHH) 人员所占比例不足。 DHH 美国人的英语阅读技能存在很大差异，他们面临着 IT 行业的挑战，在整个职业生涯中需要定期提高技能。 这项跨学科研究将重点研究自动文本简化技术作为阅读辅助，帮助 DHH 个人在正式课堂环境之外通过自主学习跟上快速变化的技术。 研究团队由自然语言处理（NLP）、人机交互（HCI）、无障碍和聋人STEM教育研究方面的专家组成。为该项目开发的资源和技术将适用于其他语言和文本领域（例如，面向非专业读者的医疗信息），使广泛的人群受益（例如，儿童、非母语人士、阅读障碍者）。针对特定用户群体的简化研究侧重于初步数据收集或经典 NLP 方法，很少有 HCI 研究探索阅读辅助系统的设计方案。该项目将填补这一研究的关键空白，并将调查提高 DHH 计算工作者技能的学习和阅读辅助需求、影响阅读辅助工具可用性的设计参数、针对特定读者群体和技术定制的新的文本简化训练数据和方法。流派、自动和基于人的评估方法，以及此类工具对自我学习的影响。这些方法将包括对 DHH 计算工作者的访谈和调查研究、设计变体的原型设计和测试、并行简化语料库的创建、DHH 个人对词典和文本的可读性注释、领域适应的 NLP 研究以及基于语法的神经机器翻译文本简化，以及与 DHH 学生和国家聋人技术学院 (NTID) 计算相关项目的应届毕业生一起观察原型的实际部署。该奖项反映了通过使用基金会的智力价值和更广泛的影响审查标准进行评估，NSF 的法定使命被认为值得支持。

项目成果

A Word-Complexity Lexicon and A Neural Readability Ranking Model for Lexical Simplification

词复杂度词典和词汇简化的神经可读性排序模型

• DOI: 10.18653/v1/d18-1410
• 发表时间: 2018-01
• 期刊: Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing
• 作者: Maddela, Mounica;Xu, Wei
• 通讯作者: Xu, Wei

BiSECT: Learning to Split and Rephrase Sentences with Bitexts

BiSECT：学习用双文本分割和改写句子

• DOI: 10.18653/v1/2021.emnlp-main.500
• 发表时间: 2021-09-10
• 期刊: ArXiv
• 作者: Joongwon Kim;Mounica Maddela;Reno Kriz;Wei Xu;Chris Callison
• 通讯作者: Chris Callison

Controllable Text Simplification with Explicit Paraphrasing

通过显式释义实现可控文本简化

• DOI: 10.18653/v1/2021.naacl-main.277
• 发表时间: 2020-10-21
• 作者: Mounica Maddela;Fern;o Alva;o;W. Xu
• 通讯作者: W. Xu

Neural semi-Markov CRF for Monolingual Word Alignment

用于单语词对齐的神经半马尔可夫条件随机场

• DOI: 10.18653/v1/2021.acl-long.531
• 发表时间: 2021-08
• 期刊: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing
• 作者: Lan, Wuwei;Jiang, Chao;Xu, Wei
• 通讯作者: Xu, Wei

Neural CRF Model for Sentence Alignment in Text Simplification

文本简化中句子对齐的神经 CRF 模型

• DOI: 10.18653/v1/2020.acl-main.709
• 发表时间: 2020-01
• 期刊: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics
• 作者: Jiang, Chao;Maddela, Mounica;Lan, Wuwei;Zhong, Yang;Xu, Wei
• 通讯作者: Xu, Wei