NSF-CNPq Collaborative Research: Advancing Human Language Technology in Brazil and the United States Through Collaborative Research on Spoken Language Systems

摘要IIS-9970061 Cole，Ronald A.&amp；Macon，Michael W.University of Colorado Boulder在99-12财年拨款99,683美元(其中49,842美元来自HCI，49,841美元来自EIA)。通过口语系统合作研究促进巴西和美国的人类语言技术这是一个为期两年的连续奖项的第一年资助。该项目需要合作研究和技术转让，从而形成巴西葡萄牙语版的CSLU语音工具包，这是一个学习、研究和开发口语系统的综合软件平台。我们将通过合作研讨会/培训课程指导来自巴西两所大学的研究人员，并合作推动口语技术的最先进水平。我们的目标是扩展工具包的能力，包括巴西葡萄牙语语音识别和文本到语音合成，并使用这些技术开发信息获取和语言培训系统。这将需要创建适当的语言资源(词典、语料库)，开发软件基础设施，并进行基础研究，以改进基本技术。然后，我们将通过向美国和巴西各地的学术机构免费分发工具包和语言资源来共享我们的知识、软件和语言资源，以促进这一研究领域的增长，并从长远来看，改善巴西和美国公民获取信息的机会。

AbstractIIS-9970061Cole, Ronald A. & Macon, Michael W. University of Colorado Boulder$99,683 ($49,842 from HCI and $49,841 from EIA) in FY99 - 12 mos.Advancing Human Language Technology in Brazil and the United States Through Collaborative Research on Spoken Language Systems This is the first year funding of a two-year continuing award. This project entails collaborative research and technology transfer leading to a Brazilian Portuguese version of the CSLU Speech Toolkit, an integrated software platform for learning about, researching and developing spoken language systems. We will mentor researchers from two universities in Brazil via collaborative workshops/training sessions and collaborate to advance the state-of-the-art of spoken language technology. Our goal is to extend the toolkit's capabilities to include Brazilian Portuguese speech recognition and text-to-speech synthesis and use these technologies to develop systems for information access and language training. This will entail creation of suitable language resources (lexicons, corpora), development of software infrastructure, and basic research to improve the underlying technology. We will then share our knowledge, software, and language resources by distributing the toolkit and language resources at no cost to academic institutions throughout the United States and Brazil to spur growth in this area of research and, in the long term, improve access to information by the citizens of both Brazil and the U.S.