Article Structure
Abstract
Written documents created through dictation differ significantly from a true verbatim transcript of the recorded speech.
Introduction
Large vocabulary speech recognition today achieves a level of accuracy that makes it useful in the production of written documents.
Text transformation
In dictation and transcription management systems corresponding pairs of recognition output and edited and corrected documents are readily available.
Transformation based learning
Following Strzalkowski and Brandow (1997) and Peters and Drexel (2004) we have implemented a transformation-based learning (TBL) algorithm (Brill, 1995).
Probabilistic model
The canonical approach to text transformation following statistical decision theory is to maximize the text document posterior probability given the spoken document.
Experimental evaluation
The methods presented were evaluated on a set of real-life medical reports dictated by 51 doctors.
Conclusions
Automatic text transformation brings speech recognition output much closer to the end result desired by the user of a back-end dictation system.
Topics
error rate
- This method iteratively improves the match (as measured by token error rate ) of a collection of corresponding source and target token sequences by positing and applying a sequence of substitution rules.Page 3, “Transformation based learning”
- The decision rule (1) minimizes the document error rate .Page 4, “Probabilistic model”
- It is hard to quote the verbatim word error rate of the recognizer, because this would require a careful and time-consuming manual transcription of the test set.Page 4, “Experimental evaluation”
- Using the alignment we compute precision and recall for sections headings and punctuation marks as well as the overall token error rate .Page 5, “Experimental evaluation”
- It should be noted that the so derived error rate is not comparable to word error rates usually reported in speech recognition research.Page 5, “Experimental evaluation”
- With 100 training documents per user the mean token error rate is reduced by up to 40% relative by the probabilistic model.Page 5, “Experimental evaluation”
See all papers in Proc. ACL 2008 that mention error rate.
See all papers in Proc. ACL that mention error rate.
Back to top.
probabilistic model
- In the case of the probabilistic model , all models were 3-gram models.Page 5, “Experimental evaluation”
- With 100 training documents per user the mean token error rate is reduced by up to 40% relative by the probabilistic model .Page 5, “Experimental evaluation”
- On the other hand the probabilistic model suffers from a slightly higher deletion rate due to being overzealous in this regard.Page 6, “Experimental evaluation”
See all papers in Proc. ACL 2008 that mention probabilistic model.
See all papers in Proc. ACL that mention probabilistic model.
Back to top.
