5,811 Hits in 7.8 sec

Using combined decisions and confidence measures for name recognition in automatic directory assistance systems

Andreas Kellner, Bernhard Rueber, Hauke Schramm
1998 5th International Conference on Spoken Language Processing (ICSLP 1998)   unpublished
Directory assistance systems are amongst the most challenging applications of speech recognition.  ...  Especially designed confidence measures for N-best lists are proposed to detect misrecognized turns.  ...  Introduction Fully automatic directory assistance is still one of the big challenges in the field of spoken dialog systems.  ... 
doi:10.21437/icslp.1998-706 fatcat:d65rfkyyybaabcqcl6bzcwn74m

Strategies for name recognition in automatic directory assistance systems

H Schramm, B Rueber, A Kellner
2000 Speech Communication  
Confidence measures are used for an optimal dialogue flow. We present results of different recognition strategies for databases of various sizes with up to 1.3 million entries (city of Berlin).  ...  The strategy is used in a prototype of a fully automated directory information system which is designed to cover a whole country: After the city has been selected, the user is asked for first and last  ...  Introduction In recent years, the challenging task of automatic directory assistance has had a lot of attention in the speech recognition community.  ... 
doi:10.1016/s0167-6393(99)00066-7 fatcat:xjhuzdwsfnhbxg7hj7sechjaia

The thoughtful elephant: strategies for spoken dialog systems

B. Souvignier, A. Kellner, B. Rueber, H. Schramm, F. Seide
2000 IEEE Transactions on Speech and Audio Processing  
They include tasks from the travel domain and automatic switchboards as well as large scale directory assistance.  ...  In particular, we demonstrate how combination decisions over several turns can be exploited to boost the recognition performance of the system.  ...  Confidence Measures for N-best Lists In view of applications where hypotheses from N-best lists are combined, one has to judge whether an N-best list is useful in a combination decision or not.  ... 
doi:10.1109/89.817453 fatcat:ijhv4wua55a27nidrthwlhynny

Utterance Verification Based on the Likelihood Distance to Alternative Paths [chapter]

Gies Bouwman, Lou Boves
2002 Lecture Notes in Computer Science  
Experiments in which different CMs were combined in different ways in the recognition of Dutch city names show that Confidence Error Rates [8] are reduced by 10% by combining the CMs using a classification  ...  For this purpose the probability of an error is often estimated by single confidence measure (CM).  ...  At the moment, we are exploring their use within the scope of the SMADA project [1] , that investigates the practical implications of developing a system for nationwide directory assistance (DA).  ... 
doi:10.1007/3-540-46154-x_29 fatcat:3noiextcbnahdmignl364j6huy

Leveraging call context information to improve confidence classification

Michael Levit
2010 2010 IEEE Spoken Language Technology Workshop  
This paper describes how speech recognition confidence estimation in a typical Directory Assistance scenario can be improved by taking dialog context into account and recalibrating the original recognition  ...  confidences using a statistical classifier that employs classification features extracted from this context.  ...  INTRODUCTION Recognition confidences have been used for many years to cope with errors in Automatic Speech Recognition.  ... 
doi:10.1109/slt.2010.5700888 dblp:conf/slt/Levit10 fatcat:iy4pgoiaq5a7ddv2sydyre7hlm

Usability of dialogue design strategies for automated surname capture

Nancie Davidson, Fergus McInnes, Mervyn A. Jack
2004 Speech Communication  
Surname capture via automatic speech recognition over the telephone has many commercial applications, including automated directory assistance and travel reservation services.  ...  The data underline the importance of user testing, demonstrating the value of the evaluation methodology used, and provide encouraging results for the strategies that involve both speaking and spelling  ...  The authors also wish to acknowledge Nuance Communications for use of their recognition software, and bmi British Midland for their contribution in providing the context for the research.  ... 
doi:10.1016/j.specom.2004.02.002 fatcat:ak3cbnakb5cl7kdjkys5xk2hee

An introduction to voice search

Ye-Yi Wang, Dong Yu, Yun-Cheng Ju, A. Acero
2008 IEEE Signal Processing Magazine  
For example, directory assistance (DA) [1] is one of the most popular voice search applications, in which users issue a spoken query and an automated system returns the phone number and address information  ...  The DM utilizes confidence measures, which indicate the certainty of the interpretations, to decide how to present the n-best results.  ...  CONFIDENCE MEASURE Confidence measures are used to determine what to do with the search results for a spoken query.  ... 
doi:10.1109/msp.2008.918411 fatcat:7h5rrhjytjcdliibeksnjpewpa

Diagnostic Decision Support System in Dysmorphology [chapter]

Kaya Kuru, Yusuf Tunc
2012 Decision Support Systems  
Acknowledgements The authors are very grateful to TÜBİTAK (The Scientific and Technological Research Council of Turkey) for sponsoring the study.  ...  In other words, all modules in Figure 3 are employed automatically. The system accepts the directory names where the dysmorphic images are as the names of the syndromes while training the datasets.  ...  Some robust composite computer assisted decision support systems are needed to be established not only for practitioners but also geneticists to support their decisions through thousands of dysmorphic  ... 
doi:10.5772/51118 fatcat:np6xdpe4pbbyjlr2camb3hy7ia

Garbage modeling with decoys for a sequential recognition scenario

Michael Levit, Shuangyu Chang, Bruce Buntschuh
2009 2009 IEEE Workshop on Automatic Speech Recognition & Understanding  
We show how this algorithm outperforms a system that only looks at recognition confidences by about 20% relative.  ...  In particular, we focus on decisions when to accept the results of the first recognizer and when the second recognizer needs to be consulted.  ...  to a better user experience, a lot of the commercial speech recognition for call centers, directory assistance and other applications still relies on directed dialog models with handcrafted, phrase-based  ... 
doi:10.1109/asru.2009.5372919 dblp:conf/asru/LevitCB09 fatcat:rgi2zdfi6vhc3iazabpsg2qb6q

Complementary computing: policies for transferring callers from dialog systems to human receptionists

Eric Horvitz, Tim Paek
2007 User modeling and user-adapted interaction  
We describe a study of the use of decision-theoretic policies for optimally joining human and automated problem-solving efforts.  ...  We describe the learning of models that predict outcomes and interaction times and show how these models can be used to generate expected-utility policies that identify when it is best to transfer callers  ...  Acknowledgments We thank Johnson Apacible for support with the data extraction and analysis and Robert Moon for assistance with gaining access to the VoiceDialer logs.  ... 
doi:10.1007/s11257-006-9026-1 fatcat:bfnattjsxjaepmn3hnq3iffghm

From Focused Crawling to Expert Information [chapter]

Sergej Sizov, Jens Graupmann, Martin Theobald
2003 Proceedings 2003 VLDB Conference  
confidence using a linear SVM) and uses them for periodically retraining the classifier; this way the crawler is dynamically adapted based on the most significant documents seen so far.  ...  for user queries, user feedback score, last modified attribute (and also combinations of these metrics).  ... 
doi:10.1016/b978-012722442-8/50116-6 dblp:conf/vldb/SizovGT03 fatcat:5etbiau55rf35isio6bcqilkm4

Learning about Voice Search for Spoken Dialogue Systems

Rebecca J. Passonneau, Susan L. Epstein, Tiziana Ligorio, Joshua B. Gordon, Pravin Bhutada
2010 North American Chapter of the Association for Computational Linguistics  
Our learned models of the best wizard's behavior combine features available to wizards with some that are not, such as recognition confidence and acoustic model scores.  ...  In a Wizard-of-Oz experiment with multiple wizard subjects, each wizard viewed automated speech recognition (ASR) results for utterances whose interpretation is critical to task success: requests for books  ...  We thank the anonymous reviewers, the Heiskell Library, our CMU collaborators, our statistical wizard Liana Epstein, and our enthusiastic undergraduate research assistants.  ... 
dblp:conf/naacl/PassonneauELGB10 fatcat:agqrliy4jzct7g2cakqgnbf45q

Confidence estimation for NLP applications

Simona Gandrabur, George Foster, Guy Lapalme
2006 ACM Transactions on Speech and Language Processing  
Confidence measures are a practical solution for improving the usefulness of Natural Language Processing applications.  ...  We give an overview of the application of confidence estimation in various fields of Natural Language Processing, and present experimental results for speech recognition, spoken language understanding,  ...  We wish to thank the members of th RALI laboratory for their support and advice, in particular Elliott Macklovitch and Philippe Langlais.  ... 
doi:10.1145/1177055.1177057 dblp:journals/tslp/GandraburFL06 fatcat:funmvci2sfhk3davwffajqkhze

Speech and language processing for next-millennium communications services

R.V. Cox, C.A. Kamm, L.R. Rabiner, J. Schroeter, J.G. Wilpon
2000 Proceedings of the IEEE  
synthesis, recognition, and understanding for dialogue access to information, people, and messaging; and speaker verification for secure access to information and services.  ...  Speech technologies have progressed to the point where they are now viable for a broad range of communications services, including compression of speech for use over wired and wireless networks; speech  ...  A measure of confidence is assigned to the recognized string, and the decision box can either accept or reject the hypothesized words by comparing the confidence scores to a decision threshold. 2) Performance  ... 
doi:10.1109/5.880086 fatcat:pzd6bawvirfo7c64t4zou2h4hy

Semi-supervised Adaptation of Assistant Based Speech Recognition Models for different Approach Areas

Matthias Kleinert, Hartmut Helmke, Gerald Siol, Heiko Ehr, Aneta Cerna, Christian Kern, Dietrich Klakow, Petr Motlicek, Youssef Oualil, Mittul Singh, Ajay Srinivasamurthy
2018 2018 IEEE/AIAA 37th Digital Avionics Systems Conference (DASC)  
MALORCA project developed an initial basic ABSR system and semi-automatically tailored its recognition models for both Prague and Vienna approach by machine learning from automatically transcribed audio  ...  Command recognition error rates were reduced from 7.9% to under 0.6% for Prague and from 18.9% to 3.2% for Vienna.  ...  ACKNOWLEDGMENT We would like to thank all the controllers who anonymously provided us with real world command examples and also our MALORCA partners from Austro Control and from Air Navigation Service  ... 
doi:10.1109/dasc.2018.8569879 fatcat:p52kdrjjy5a5pijoqzo2avcfvu
« Previous Showing results 1 — 15 out of 5,811 results