Noun-phrase co-occurrence statistics for semiautomatic semantic lexicon construction
COLING '98 Proceedings of the 17th international conference on Computational linguistics - Volume 2
Speech-based retrieval using semantic co-occurrence filtering
HLT '94 Proceedings of the workshop on Human Language Technology
Discovering Cues to Error Detection in Speech Recognition Output: A User-Centered Approach
Journal of Management Information Systems
Enhancing accessibility through correction of speech recognition errors
ACM SIGACCESS Accessibility and Computing - ASSETS 2007 doctoral consortium
A Hybrid Approach to Improving Automatic Speech Recognition Via NLP
CAI '07 Proceedings of the 20th conference of the Canadian Society for Computational Studies of Intelligence on Advances in Artificial Intelligence
IWSDS'10 Proceedings of the Second international conference on Spoken dialogue systems for ambient environments
The effect of noise in automatic text classification
Proceedings of the International Conference & Workshop on Emerging Trends in Technology
Detection of semantic errors in Arabic texts
Artificial Intelligence
An illustrated methodology for evaluating ASR systems
AMR'11 Proceedings of the 9th international conference on Adaptive Multimedia Retrieval: large-scale multimedia retrieval and evaluation
Information distance between what I said and what it heard
Communications of the ACM
Statistical error correction methods for domain-specific ASR systems
SLSP'13 Proceedings of the First international conference on Statistical Language and Speech Processing
Hi-index | 0.02 |
In this paper we present preliminary results of a novel unsupervised approach for high-precision detection and correction of errors in the output of automatic speech recognition systems. We model the likely contexts of all words in an ASR system vocabulary by performing a lexical co-occurrence analysis using a large corpus of output from the speech system. We then identify regions in the data that contain likely contexts for a given query word. Finally, we detect words or sequences of words in the contextual regions that are unlikely to appear in the context and that are phonetically similar to the query word. Initial experiments indicate that this technique can produce high-precision targeted detection and correction of misrecognized query words.