Paper: Automatic Spoken Document Processing For Retrieval And Browsing

ACL ID N06-5004
Title Automatic Spoken Document Processing For Retrieval And Browsing
Venue Human Language Technologies
Session Tutorial Abstracts
Year 2006
Authors

* Tuning precision/recall at query run-time – Evaluation: * Basic Metrics: Precision/Recall * Ordered list metrics: Kendall-Tau, Spearman * TREC measures and package (Mean Average Precision, R-precision) * Issues with evaluating speech data – User interface: * Issues in consuming speech (as opposed to text, images) * Pro’s/con’s for displaying transcription * Navigation in long documents with errorful transcriptions * Segmentation: topic boundaries, keywords, summaries 4.2 Target Audience This tutorial is designed for people interested in learning about the technologies used to transcribe, process, search, and retreive spoken audio materials. Detailed prior knowledge of speech recognition and/or search technologies is not required. Ciprian Chelba is a Research Scientist with Google...