Paper: Score Distribution Based Term Specific Thresholding for Spoken Term Detection

ACL ID N09-2068
Title Score Distribution Based Term Specific Thresholding for Spoken Term Detection
Venue Human Language Technologies
Session Short Paper
Year 2009
Authors

The spoken term detection (STD) task aims to return relevant segments from a spoken archive that contain the query terms. This pa- per focuses on the decision stage of an STD system. We propose a term specific threshold- ing (TST) method that uses per query poste- rior score distributions. The STD system de- scribed in this paper indexes word-level lat- tices produced by an LVCSR system using Weighted Finite State Transducers (WFSTs). The target application is a sign dictionary where precision is more important than recall. Experiments compare the performance of dif- ferent thresholding techniques. The proposed approach increases the maximum precision at- tainable by the system.