2003 Symposium on Next Generation Automatic Speech Recognition (ASR)

尽管在过去的二十年中，使用模式匹配范式的自动语音识别（ASR）取得了重大进展，但最近进展开始放缓。此外，目前最先进的技术还无法与人类语音识别能力相媲美。一个潜在的改进领域是找到评估和整合ASR社区之外开发的大量知识来源的方法，并建立开放的共享基础设施，以促进ASR合作研究。由于语音识别技术对于实现所有人普遍获取信息的承诺至关重要，因此研究社区可以采取的方案以弥合性能差距是及时的。该项目将组织一个高度集中的为期两天的研讨会，邀请美国学术界、工业界和政府的顶尖人士评估最先进的ASR技术的能力和局限性，并讨论加速ASR进展的潜在方法。研讨会的主题将集中于整合多学科的知识来源，从声学、语音、语言学、人机交互、认知科学、信号处理和计算机科学，到ASR组件和系统设计的每个阶段。在会议结束时，所有的演示材料将被整理成一份报告，并分发给研究界和其他利益相关者。这项研究将激发ASR研究的新方向，为语音科学和技术的研究和教育构想新的网络基础设施，并通过建立多元化的研究人员社区，为降低ASR的进入门槛和增加人类语言和交流研究的人才库提供建议。

Although there have been significant advances in automatic speech recognition (ASR) over the last twenty years using a pattern-matching paradigm, the progress has begun to slow recently. Furthermore, the current state-of-the-art has yet to rival human speech recognition capability. One potential area for improvement is to find ways to evaluate and incorporate the vast knowledge sources developed outside the ASR community and establish open sharable infrastructures to facilitate collaborative ASR research. As speech recognition technology is critical to realize the promise of universal access to information for all people, it is timely to examine options the community can take up to bridge the performance gap.This project is to organize a highly focused two-day symposium and invite top minds in US academia, industry and government to assess the capabilities and limitations of state-of-the-art ASR technologies, and discuss potential approaches to accelerating ASR advances. The symposium theme will be centered on integrating multi-disciplinary sources of knowledge, from acoustics, speech, linguistics, human computer interaction, cognitive science, signal processing, and computer science, into every stage of ASR component and system design. At the conclusion of the meeting, all the presentation materials will be collated into a report and disseminated to the research community and other stakeholders. This research should inspire new directions in ASR, conceptualize new cyber-infrastructures for research and education in speech science and technology, and recommend ways to lower ASR entry barriers and increase the talent pool in human language and communication research by enabling a diverse community of researchers.