... 764–772,Suntec, Singapore, 2-7 August 2009.c2009 ACL and AFNLP Improving AutomaticSpeechRecognitionforLectures throughTransformation-based Rules Learned from Minimal DataCosmin Munteanu∗†∗National ... scoring function.1 Introduction Improving access to archives of recorded lectures is a task that, by its very nature, requires researchefforts common to both AutomaticSpeech Recog-nition (ASR) ... Workshopon AutomaticSpeechRecognition and Understand-ing (ASRU), pages 347–354.C. F¨ugen, M. Kolss, D. Bernreuther, M. Paulik,S. St¨uker, S. Vogel, and A. Waibel. 2006. Opendomain speech recognition...
... this paper the best performing measuresfrom (Pucher, 2005), which outperform baselinemodels on word prediction for conversational tele-phone speech are used forAutomaticSpeech Recog-nition ... 129–132,Prague, June 2007.c2007 Association for Computational LinguisticsWordNet-based Semantic Relatedness Measures in Automatic Speech Recognitionfor MeetingsMichael PucherTelecommunications ... conversational speech. The JCN (Sec-tion 2.1) measure performs best for nouns using thenoun-context. The LESK (Section 2.1) measure per-forms best for verbs and adjectives using a mixedword-context.Text-based...
... transcription. For example, if the ASR output contains the term sequence “… and farther home run for David forty says…” and the closed captioning contains the sequence “…another home run for David ... evaluate the performance of our grounded language model on a speechrecognition task using video highlights from Major League Baseball games. Results indicate improved per-formance using three ... captioning transcrip-tions (i.e., no ASR). 5 Conclusions We have described a method forimprovingspeech recognition in video. The method uses grounded language modeling, an extension of tradition...
... for spontaneous speech recognition One of the most important issues forspeech recognition is how to create language models (rules) for spontaneous speech. When recognizing spontaneous speech ... travel information system at LIMSI The ARISE (Automatic Railway Information Systems for Europe) projects aims developing prototype telephone information services for train travel information ... ROBUST SPEECH RECOGNITION 4.1 Automatic adaptation Ultimately, speech recognition systems should be capable of f robust, speaker- independent or speaker- adaptive, continuous speech recognition ...
... multiple speech recognizers in an effort to improve speech recognition, as discussed next. 2.1 Enhanced Majority Rules Barry (et al., 1994) took three different Automatic Speech Recognition ... the same input, performed speech recognition, and sent the result to the master system. The EMR resolved inconsistencies by looking for agreement from the individual systems for the recognized ... & Gauvain, J., Improved ROVER using Language Model Information, In ISCA ITRW Workshop on AutomaticSpeech Recognition: Chal-lenges for the new Millenium, Paris, pp. 47–52, 2000. Young,...
... hybrid sys-tem for open vocabulary speech recognition. Ratherthan relying on the text alone, we also utilize sideinformation: a mapping of words to classes so wecan optimize learning for a specific ... and MIT Lectures task respectively.1 IntroductionMost automaticspeechrecognition systems operatewith a large but limited vocabulary, finding the mostlikely words in the vocabulary for the ... b ax,d ae n. The latter is more useful for automaticallyrecovering the word’s orthographic form, identify-ing that an OOV was spoken, or improving perfor-mance of a spoken term detection system...
... Previous WorkTechniques for exploiting stochastic context-freegrammars for language modeling have been ex-plored for more than a decade. Early approachesincluded algorithms for efficiently calculating ... are a first step inexamining the potential utility of syntactic features for discriminative language modeling for speech recognition. We tried two possible sets of featuresderived from the full ... Using a stochastic context-free grammar as a lan-guage model forspeech recognition. In Proceedings of theIEEE Conference on Acoustics, Speech, and Signal Process-ing, pages 189–192.John Lafferty,...
... to correct the errors in the results of speech recognition to increase the performance of a speech translation system. This paper proposes a method for correcting errors using the statistical ... integrating recognition and translation into a speech translation system, the development of the following processes is therefore important: (1) detection of errors in speech recognition results; ... correct string for the string between A and B in the Error-String (see figure 2-3). 3. Evaluation 3.1 Data Condition for Experiments Results of Speech Recognition: We used 4806 recognition...
... invariants; allophonic variation is traditionally seen as problematic for recognition. (I) "In most systems for sentence recognition, such modifications must be viewed as a kind of 'noise' ... before the labial stop /p/, the cor9nal nasal/n/ before the coronal stop/t/, and the velar nasal/7// before the velar stop/k/. This constraint, like subject-verb agreement. poses a problem for ... This view of allophonic variation is representative of much of the speech recognition literature, especially during the ARPA speech project. One can find similar statements by Cole and Jakim~k...
... rates were 17.0 for En-glish, 18.7 for Spanish, and 22.5 for French. For English, we also created web mixture mod-els with KN smoothing. The error rates were 16.5,15.9 and 15.7 for the 20 MB, ... 2.2.1) for the same number of queries. Alsoresults from language modeling and speech recog-nition experiments favored statistical querying.2.3 Web collections obtained For the speechrecognition ... room for improvement.3.1.2 Word error rates Speech recognition results for the different LMsare given in Table 2. The results are consistent inthe sense that the web mixture models outperformthe...
... sizes.1 Introduction1.1 OOV problemOpen vocabulary speechrecognition refers to au-tomatic speechrecognition (ASR) of continuous speech, or speech- to-text” of spoken language,where the recognizer ... Computational Model for Word-Form Recognition and Production. University of Helsinki,Helsinki, Finland.Tanel Alumäe. 2006. Methods for Estonian Large Vo-cabulary Speech Recognition. PhD Thesis. ... Vocab-ulary SpeechRecognition with Flat Hybrid Models.INTERSPEECH-2005, 725–728.Janne Pylkkönen. 2005. An Efficient One-pass Decoder for Finnish Large Vocabulary Continuous Speech Recognition. ...
... developed for the preference-first parsing algorithm for different applications. For example, there can be various construction principles to determine the order of constituent construction for ... Unification Grammar and Markov Language Model for Continuous Speech Recognition. Proceedings of the IEEE 990 International Conference on Acoustics, Speech and Signal Processing, Albuquerque, NM, ... (1986). An Efficient Word Lattice Parsing Algorithm for Continuous Speech Recognition. Proceedings of the 1986 International Conference on Acoustic, Speech and Signal Processing, pp. 1569-1572....
... Spontaneous Speech in Unrestricted Domains for Reducing Word Error Rate in SpeechRecognition Klaus Zechner and Alex Waibel Language Technologies Institute Carnegie Mellon University 5000 Forbes ... strong rationale for following this simple ap- proach is the nature of the ill-formed input due to (i) spontaneous speech dysfluencies, and (ii) errors in the hypotheses of the speech recognizer. ... lattices. The hypotheses from the Nbest lists are tagged for part of speech, "cleaned up" by a preprocessing pipe, parsed by a part of speech based chunk parser, and rescored using a backpropagation...