1 / 13

Rapid and Accurate Spoken Term Detection

Rapid and Accurate Spoken Term Detection. Owen Kimball BBN Technologies 15 December 2006. Overview of Talk. BBN Levantine system description Evaluation results Diacritics Out-of-vocabulary issues. Additional assistance Thomas Colthurst Herb Gish Steve Lowe Rich Schwartz.

barrett
Download Presentation

Rapid and Accurate Spoken Term Detection

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Rapid and Accurate Spoken Term Detection Owen Kimball BBN Technologies 15 December 2006

  2. Overview of Talk • BBN Levantine system description • Evaluation results • Diacritics • Out-of-vocabulary issues Rapid and Accurate Spoken Term Detection

  3. Additional assistance Thomas Colthurst Herb Gish Steve Lowe Rich Schwartz BBN Evaluation Team Core Team • Chia-lin Kao • Owen Kimball • Michael Kleber • David Miller Rapid and Accurate Spoken Term Detection

  4. BBN System Overview audio search terms ATWV cost parameters Byblos STT indexer lattices phonetic- transcripts detector index decider scored detection lists final output with YES/NO decisions Rapid and Accurate Spoken Term Detection

  5. Levantine STT Configuration • STT generates a lattice of hypotheses and a phonetic transcript for each input file. • Word-based system: • Orthography based on Modern Standard Arabic (MSA), no short vowel diacritics • Acoustic: 57.3 hours LDC (noise words, no mixture exponents) • Language: 250 hours of data, 1.3M words • 38.5K dictionary, grapheme-as-phoneme based plus 100 manual pronunciations • unknown short vowel (U), 39 phonemes • 42.32% WER on STD Dev06 CTS data Rapid and Accurate Spoken Term Detection

  6. Levantine CTS Results Data ATWV Dev06 0.515 DryRun 0.410 Eval06 0.3467 Rapid and Accurate Spoken Term Detection

  7. OOV Pipeline: Detector • Word-based STT produces 1-best transcript: pronounce it  1-best phonetic transcript. • Query is OOV if it contains any OOV word. • OOV query detection: • Pronounce query (grapheme-as-phoneme) • Find minimal edit-distance alignments (agrep) • Score = % error = Rapid and Accurate Spoken Term Detection

  8. OOV Pipeline: Decider • Need different Yes/No decision procedure:IV-decider requires posterior probabilities. • Simple OOV decision procedure: • Constant threshold on score (~ 0.7) • Cap on maximum number of hits (0-3) • Values set to maximize ATWV on Dev06 data. Rapid and Accurate Spoken Term Detection

  9. OOV Pipeline: Results • ATWV remained good: 0.3450 IV 0.3635 OOV • Searches take longer: ~10-15x IV speed on Dev06 and DryRun06, with no attempt at indexing. Rapid and Accurate Spoken Term Detection

  10. OOV Directions for Improvement • Score substitutions using phoneme confusion matrix instead of flat edit distance • Speed: indexing phonetic transcripts for approximate matching • Search lattices beyond 1-best transcripts Rapid and Accurate Spoken Term Detection

  11. Levantine Diacritic Issues • Originally looked at diacritized Levantine • Trained STT engine using LDC 45 hour set • Ran STD without knowing WER (no diacritized STT test set to measure WER). • Found very high false alarm rate • Examining FAs found hits that were legitimate alternate spellings Rapid and Accurate Spoken Term Detection

  12. Levantine Diacritics- Alternate Spellings • Examining query words found more of same: • In first 22 terms of dry run term list, 14 are “alternate diacritic” spellings of 5 underlying words, i.e. there were just 13 unique words in the first 22 terms • Min~ahumo v Minohumo • AlHayaApi v AlHayaAp • Waliko v Walika • qabilo v qabola v qabolo • LDC training and STD test set had additional pervasive differences Rapid and Accurate Spoken Term Detection

  13. No-Diacritic Levantine Issues • A quick look turned up a smaller number of problems for no-diacritic Levantine • Looking at 7 top-FA terms in dev set, found • “bHky” vs “b>Hky” but no other spelling confusions • One ref instance of term with 0 duration • It would be interesting to QC test sets for inconsistent spellings and other issues Rapid and Accurate Spoken Term Detection

More Related