TY - GEN

T1 - Support vector machine re-scoring of hidden Markov models

AU - Sloin, Alba

AU - Burshtein, David

PY - 2006

Y1 - 2006

N2 - We present a method that uses a set of maximum-likelihood (ML) trained discrete HMM models as a baseline system, and an SVM training scheme to re-score the results of the baseline HMMs. It turns out that the re-scoring model can be represented as an un-normalized HMM. We refer to these models as pseudo-HMMs. The pseudo-HMMs are in fact a generalization of standard HMMs, and by proper discriminative training they can result in performance improvement compared to standard HMMs. We consider two SVM training algorithms. The first corresponds to the one against all method. The second corresponds to the one class transformation training method. The one class training algorithm can be extended to an iterative algorithm, similar to segmental K-means. In this case the final output of the algorithm is a single set of pseudo-HMMs. Although they are not normalized, this set of pseudo-HMMs can be used in the standard recognition procedure (the Viterbi recognizer), as if they were plain HMMs. We use an isolated noisy digit recognition task to demonstrate that SVM re-scoring of HMMs typically reduces the error rate significantly compared to standard ML training.

AB - We present a method that uses a set of maximum-likelihood (ML) trained discrete HMM models as a baseline system, and an SVM training scheme to re-score the results of the baseline HMMs. It turns out that the re-scoring model can be represented as an un-normalized HMM. We refer to these models as pseudo-HMMs. The pseudo-HMMs are in fact a generalization of standard HMMs, and by proper discriminative training they can result in performance improvement compared to standard HMMs. We consider two SVM training algorithms. The first corresponds to the one against all method. The second corresponds to the one class transformation training method. The one class training algorithm can be extended to an iterative algorithm, similar to segmental K-means. In this case the final output of the algorithm is a single set of pseudo-HMMs. Although they are not normalized, this set of pseudo-HMMs can be used in the standard recognition procedure (the Viterbi recognizer), as if they were plain HMMs. We use an isolated noisy digit recognition task to demonstrate that SVM re-scoring of HMMs typically reduces the error rate significantly compared to standard ML training.

UR - http://www.scopus.com/inward/record.url?scp=46749111477&partnerID=8YFLogxK

U2 - 10.1109/EEEI.2006.321107

DO - 10.1109/EEEI.2006.321107

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:46749111477

SN - 1424402301

SN - 9781424402304

T3 - IEEE Convention of Electrical and Electronics Engineers in Israel, Proceedings

SP - 376

EP - 380

BT - 2006 IEEE 24th Convention of Electrical and Electronics Engineers in Israel, IEEEI

T2 - 2006 IEEE 24th Convention of Electrical and Electronics Engineers in Israel, IEEEI

Y2 - 15 November 2006 through 17 November 2006

ER -