RIA: Vocabulary-Independent Modeling for Continuous Speech Recogntion

RIA：连续语音识别的词汇无关建模

• 批准号: 9010432
• 负责人: Kai-Fu Lee
• 金额: $ 5.88万
• 依托单位: Carnegie-Mellon University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1990
• 资助国家: 美国
• 起止时间: 1990-07-01 至 1992-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9010432&HistoricalAwards=false
• 关键词: RIA; Vocabulary; Independent; Modeling; Continuous

This Research Initiation Award will help Dr. Lee to extend the highly successful SPHINX speech recognition system to require less vocabulary-specific training for new tasks. Proposed techniques include linguistic modeling of allophonic variations, automatic subword clustering, and incremental vocabulary adaptation. The research should provide insights into subword modeling and fusion of linguistic theory, statistical learning, and incremental learning theory.

该研究启动奖将帮助 Lee 博士扩展非常成功的 SPHINX 语音识别系统，以减少新任务所需的特定词汇训练。 提出的技术包括异音变体的语言建模、自动子词聚类和增量词汇适应。 该研究应该提供对子词建模以及语言理论、统计学习和增量学习理论融合的见解。