Results 1 - 10 of 387

A.: Interpreting bleu/nist scores: How much improvement do we need to have a better system. In

by Ying Zhang, Stephan Vogel, Alex Waibel - In Proceedings of Proceedings of Language Resources and Evaluation (LREC-2004. (2004) 2051–2054 Callison-Burch , 2012
"... Automatic evaluation metrics for Machine Translation (MT) systems, such as BLEU and the related NIST metric, are becoming increasingly important in MT. Yet, their behaviors are not fully understood. In this paper, we analyze some flaws in the BLEU/NIST metrics. With a better understanding of these p ..."
Abstract - Cited by 43 (4 self) - Add to MetaCart
of these problems, we can better interpret the reported BLEU/NIST scores. In addition, this paper reports a novel method of calculating the confidence intervals for BLEU/NIST scores using bootstrapping. With this method, we can determine whether two MT systems are significantly different from each other.

The Alignment Template Approach to Statistical Machine Translation

by Franz Josef Och, Hermann Ney , 2004
"... A phrase-based statistical machine translation approach — the alignment template approach — is described. This translation approach allows for general many-to-many relations between words. Thereby, the context of words is taken into account in the translation model, and local changes in word order f ..."
Abstract - Cited by 480 (26 self) - Add to MetaCart
–English Canadian Hansards task, the alignment template system obtains significantly better results than a single-word-based translation model. In the Chinese–English 2002 National Institute of Standards and Technology (NIST) machine translation evaluation it yields statistically significantly better NIST scores

Speaker verification using Adapted Gaussian mixture models

by Douglas A. Reynolds, Thomas F. Quatieri, Robert B. Dunn - Digital Signal Processing , 2000
"... In this paper we describe the major elements of MIT Lincoln Laboratory’s Gaussian mixture model (GMM)-based speaker verification system used successfully in several NIST Speaker Recognition Evaluations (SREs). The system is built around the likelihood ratio test for verification, using simple but ef ..."
Abstract - Cited by 1010 (42 self) - Add to MetaCart
In this paper we describe the major elements of MIT Lincoln Laboratory’s Gaussian mixture model (GMM)-based speaker verification system used successfully in several NIST Speaker Recognition Evaluations (SREs). The system is built around the likelihood ratio test for verification, using simple

NIST Scoring Package Certification Procedures in Conjunction with NIST Special Databases 2 and 6

by Michael D. Garris
"... this document have been developed in conjunction with NIST Special Database 2 ..."
Abstract - Add to MetaCart
this document have been developed in conjunction with NIST Special Database 2

A Post-Processing System To Yield Reduced Word Error Rates: Recognizer Output Voting Error Reduction (ROVER)

by Jonathan G. Fiscus , 1997
"... This paper describes a system developed at NIST to produce a composite Automatic Speech Recognition (ASR) system output when the outputs of multiple ASR systems are available, and for which, in many cases, the composite ASR output has lower error rate than any of the individual systems. The system i ..."
Abstract - Cited by 422 (2 self) - Add to MetaCart
This paper describes a system developed at NIST to produce a composite Automatic Speech Recognition (ASR) system output when the outputs of multiple ASR systems are available, and for which, in many cases, the composite ASR output has lower error rate than any of the individual systems. The system

Front End Factor Analysis for Speaker Verification

by Najim Dehak, Patrick J. Kenny, Réda Dehak, Pierre Dumouchel, Pierre Ouellet - IEEE Transactions on Audio, Speech and Language Processing , 2010
"... Abstract—This paper presents an extension of our previous work which proposes a new speaker representation for speaker verification. In this modeling, a new low-dimensional speaker- and channel-dependent space is defined using a simple factor analysis. This space is named the total variability space ..."
Abstract - Cited by 315 (22 self) - Add to MetaCart
results are obtained when LDA is followed by WCCN. We achieved an equal error rate (EER) of 1.12 % and MinDCF of 0.0094 using the cosine distance scoring on the male English trials of the core condition of the NIST 2008 Speaker Recognition Evaluation dataset. We also obtained 4 % absolute EER improvement

The TREC spoken document retrieval track: A successful story

by John S. Garofolo, Cedric G. P. Auzanne, Ellen M. Voorhees - In Proceedings of the Nineth Text REtrieval Conference (TREC-9 , 2000
"... This paper describes work within the NIST Text REtrieval Conference (TREC) over the last three years in designing and implementing evaluations of Spoken Document Retrieval (SDR) technology within a broadcast news domain. SDR involves the search and retrieval of excerpts from spoken audio recordings ..."
Abstract - Cited by 174 (2 self) - Add to MetaCart
This paper describes work within the NIST Text REtrieval Conference (TREC) over the last three years in designing and implementing evaluations of Spoken Document Retrieval (SDR) technology within a broadcast news domain. SDR involves the search and retrieval of excerpts from spoken audio recordings

Cosine Similarity Scoring without Score Normalization Techniques

by Najim Dehak, Reda Dehak, James Glass, Douglas Reynolds, Patrick Kenny
"... In recent work [1], a simplified and highly effective approach to speaker recognition based on the cosine similarity between lowdimensional vectors, termed ivectors, defined in a total variability space was introduced. The total variability space representation is motivated by the popular Joint Fact ..."
Abstract - Cited by 36 (3 self) - Add to MetaCart
Factor Analysis (JFA) approach, but does not require the complication of estimating separate speaker and channel spaces and has been shown to be less dependent on score normalization procedures, such as z-norm and t-norm. In this paper, we introduce a modification to the cosine similarity that does

THE HKCUPU SYSTEM FOR THE NIST 2010 SPEAKER RECOGNITION EVALUATION

by Weiwu Jiang, Man-wai Mak, Wei Rao, Helen Meng
"... This paper presents the HKCUPU speaker recognition system submitted to NIST 2010 speaker recognition evaluation (SRE). The system comprises five subsystems, each with different acoustic features, session-variability reduction methods, speaker modeling and scoring methods and classifiers. This paper ..."
Abstract - Cited by 1 (1 self) - Add to MetaCart
This paper presents the HKCUPU speaker recognition system submitted to NIST 2010 speaker recognition evaluation (SRE). The system comprises five subsystems, each with different acoustic features, session-variability reduction methods, speaker modeling and scoring methods and classifiers. This paper

APPLICATION OF NEW I-VECTOR CONDITIONING ALGORITHM AND SCORING METHOD TO NIST SPEAKER RECOGNITION EVALUATION 2010

by Pierre-michel Bousquet, Anthony Larcher, Driss Matrouf, Bin Ma
"... Recently developed Total Variability space framework has be-come a standard for speaker verification engines. Represen-tations of speech segments in this low dimensionality space, called i-vectors, offer new opportunities to carry out speaker recognition. The nature of i-vectors requires new interse ..."
Abstract - Add to MetaCart
those techniques, we present new results and show their abil-ities to perform channel-blind compensation. In order to deal with the highly demanding cost function defined for NIST SRE 2010, we introduce a new reliability indicator which is shown to reduce false alarms by minimizing high impostor scores.
Results 1 - 10 of 387