Workshops on NASA Apollo Mission Audio as a Community Research Resource

将 NASA 阿波罗任务音频作为社区研究资源的研讨会

• 批准号: 1943365
• 负责人: John Hansen
• 金额: $ 3.09万
• 依托单位: University of Texas at Dallas
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-09-01 至 2020-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1943365&HistoricalAwards=false
• 关键词: Workshops; NASA; Apollo; Mission; Audio

The ability for human teams to work collaboratively together to solve complex problems relies on effective communications, typically relying on vocal communications. Unfortunately, the availability of naturalistic team/task-oriented communications for research in speech and language processing, as well as human factors and psychology for problem solving and team dynamics does not exist today. This project will involve the organization of a series of mini-workshops to advance the development of a future community resource based on the first massively large corpus of naturalistic data spanning years from hundreds of individuals working on Apollo missions, which solved one of humanity's greatest challenges -- placing a man on the moon and returning him to Earth safely. This future community resource will benefit research communities spanning technology, science, education, and history of technology. A clear added benefit is the preservation of the voices of 1000's of scientists, engineers, and technicians who made one of humanity's greatest technical achievements possible. This project will conduct mini-workshops to collect research community input for the development of the first massively large community resource of naturalistic audio data spanning years from hundreds of individuals working on Apollo missions. This corpus, called FEARLESS STEPS, was created as a result of a previous NSF award and represents a real-world, multi-subject, problem-solving communications corpus that is real and not simulated. The collected feedback from these mini-workshops will inform the team that created the original corpus of the community users' current needs and requirements for the future community resource. These mini-workshops would include events organized at: (i) ISCA INTERSPEECH-19, (ii) U.S. Library of Congress (Audio Preservation Day), (iii) ASA-19 Acoustical Society, (iv) NASA Human Performance Meeting, and (v) Education/Historian visits. The technical goals of the mini-workshops are: (i) conducting community outreach mini-workshops to collect feedback from the science and technology communities, as well as those in historical and educational domains; (ii) developing a platform for collecting and analyzing feedback from the research community; and (iii) providing infrastructure support to develop example FEARLESS STEPS Apollo audio sets to be shared at workshop/meetings for continued community outreach. Research advancements today in speech and language technology rest on advanced machine learning concepts, which require extensive audio data for training effective models, and it is imperative that the academic community have access to massive audio data which is naturalistic, real-world, multi-speaker, task directed, and freely available to all.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.

人类团队协作解决复杂问题的能力依赖于有效的沟通，通常依赖于声音沟通。不幸的是，今天不存在用于语音和语言处理研究的自然主义团队/面向任务的通信，以及用于解决问题和团队动态的人为因素和心理学。该项目将组织一系列小型研讨会，以推进未来社区资源的开发，该资源基于来自参与阿波罗任务的数百名个人的第一个大规模自然数据语料库，这些数据跨越数年，解决了人类面临的最大挑战之一——将人送上月球并安全返回地球。这一未来的社区资源将使技术、科学、教育和技术史领域的研究社区受益。一个明显的额外好处是保存了数千名科学家、工程师和技术人员的声音，正是他们使人类最伟大的技术成就之一成为可能。该项目将举办小型研讨会，收集研究社区的意见，以开发第一个大规模的自然音频数据社区资源，这些数据来自数百名参与阿波罗任务的个人，历时数年。该语料库名为“FEARLESS STEPS”，是之前获得 NSF 奖项而创建的，代表了真实世界、多主题、解决问题的通信语料库，是真实的而非模拟的。从这些小型研讨会收集的反馈将告知创建原始语料库的团队社区用户当前的需求以及对未来社区资源的要求。这些小型研讨会将包括以下组织的活动：(i) ISCA INTERSPEECH-19、(ii) 美国国会图书馆（音频保护日）、(iii) ASA-19 声学协会、(iv) NASA 人类表现会议和 (v) 教育/历史学家访问。小型研讨会的技术目标是： (i) 举办社区外展小型研讨会，收集科技界以及历史和教育领域的反馈； (ii) 开发一个收集和分析研究界反馈的平台； (iii) 提供基础设施支持，以开发示例 FEARLESS STEPS Apollo 音频集，以便在研讨会/会议上共享，以持续进行社区外展。当今语音和语言技术的研究进步依赖于先进的机器学习概念，这需要大量的音频数据来训练有效的模型，学术界必须能够获得大量自然、真实、多说话者、任务导向且免费提供给所有人的音频数据。该奖项反映了 NSF 的法定使命，并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，被认为值得支持。

项目成果

The 2019 Inaugural Fearless Steps Challenge: A Giant Leap for Naturalistic Audio

• DOI: 10.21437/interspeech.2019-2301
• 发表时间: 2019-09
• 作者: J. Hansen;Aditya Sunil Joglekar;M. C. Shekhar;Vinay Kothapally;Chengzhu Yu;L. Kaushik;A. Sangwan