Beginning in May 2002, the speech group at ICSI began work on the DARPA Effective, Affordable Reusable Speech-to-text (EARS) program. The goal of this five-year project is to significantly advance the state of the art in multi-lingual speech recognition of both broadcast news and human-to-human conversational speech. The EARS program consists of two subprojects: Rich Transcription and Novel Approaches. ICSI is a team member on the Rich Transcription project and the lead site for the Novel Approaches project.

Team Members for EARS

Rich Transcription

Novel Approaches

SRI (lead) ICSI (lead)
ICSI Columbia University
Univ. of Washington IDIAP
Univ. of Washington

ICSI researchers working on the EARS projects are Barry Chen, George Doddington, Yan Huang, Nelson Morgan, Barbara Peskin, Elizabeth Shriberg (also at SRI), Andreas Stolcke (also at SRI), Chuck Wooters, and Qifeng Zhu. Visiting researchers Yang Liu and Jitendra Ajmera are also currently working on the project.

You can find relevant publications and presentationshere.