Language model combination and adaptation using weighted finite state transducers

In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaptation may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences. ©2010 IEEE.

Estudios de Historia de España, 2008, vol. 10 (número completo)

Contenido: El regadío sostenible en al-Andalus / José Roldán Cañas ; María Fátima Moreno Pérez ; José Luis del Pino García – Presencia jurídica femenina a través de los Ordenamientos de Cortes (Castilla – León, Siglos XII-XIV) / Diana Arauz Mercado – Santo Domingo de Caleruega. Estudio iconológico / María Jesús Baquero Martín – Un ejemplo de poder monástico femenino: las relaciones entre la villa de Tordesillas y el convento de Santa Clara / Cecilia Bahr – La exposición de la doctrina de la “Guerra Justa” en El Victorial de Gutierre Díaz de Games / Santiago Agustín Pérez – Algunas notas sobre el patrimonio de la boticaria Elvira Pérez (Santiago, 1348) / Carlos Calderón – The price of Charles V’S protection in Italy: the example of Lucca / Christine Shaw – El Imperio y las Cortes de Santiago de Compostela de 1520, la base ideológica del Absolutismo español / István Szászdi – The extradition treaties of the Spanish and Portuguese Inquisitions (1500-1700) / François Soyer – Las “Memorias” del general argentino Tomás de Iriarte sobre la Guerra de la Independencia Española / Miguel Ángel De Marco – Los viajes a Alemania de Ramiro de Maeztu / Ángeles Castro Montero -- Reseñas

Improving speech transcription for Mandarin-english translation

This paper describes the development of the CU-HTK Mandarin Speech-To-Text (STT) system and assesses its performance as part of a transcription-translation pipeline which converts broadcast Mandarin audio into English text. Recent improvements to the STT system are described and these give Character Error Rate (CER) gains of 14.3% absolute for a Broadcast Conversation (BC) task and 5.1% absolute for a Broadcast News (BN) task. The output of these STT systems is then post-processed, so that it consists of sentence-like segments, and translated into English text using a Statistical Machine Translation (SMT) system. The performance of the transcription-translation pipeline is evaluated using the Translation Edit Rate (TER) and BLEU metrics. It is shown that improving both the STT system and the post-STT segmentations can lower the TER scores by up to 5.3% absolute and increase the BLEU scores by up to 2.7% absolute. © 2007 IEEE.

The CU-HTK Mandarin broadcast news transcription system

This paper discusses the development of the CU-HTK Mandarin Broadcast News (BN) transcription system. The Mandarin BN task includes a significant amount of English data. Hence techniques have been investigated to allow the same system to handle both Mandarin and English by augmenting the Mandarin training sets with English acoustic and language model training data. A range of acoustic models were built including models based on Gaussianised features, speaker adaptive training and feature-space MPE. A multi-branch system architecture is described in which multiple acoustic model types, alternate phone sets and segmentations can be used in a system combination framework to generate the final output. The final system shows state-of-the-art performance over a range of test sets. ©2006 British Crown Copyright.

Representação parlamentar dos Estados

Consultoria Legislativa - Área I - Direito Constitucional, Eleitoral, Municipal, Administrativo, Processo Legislativo e Poder Judiciário.

Extensão de benefícios aos portadores de narcolepsia

Avalia a conveniência de incluir a narcolepsia no rol das doenças cujos portadores usufruem de benefícios especificados em lei. Analisa também a regulamentação jurídica sobre o assunto.

Report of the 2004 Workshop on In Situ Iron Enrichment Experiments in the Eastern and Western Subarctic Pacific

Foreword 1. BACKGROUND AND OBJECTIVES (pdf, 0.1 Mb) 2. 2004 WORKSHOP SUMMARY (pdf, < 0.1 Mb) 2.1. What have we learned from the enrichment experiments? 2.2 What are the outstanding questions? 2.3 Recommendations for SEEDS-II 3. EXTENDED ABSTRACTS OF THE 2004 WORKSHOP 3.1 Synthesis of the Iron Enrichment Experiments: SEEDS and SERIES (pdf, 0.5 Mb) Iron fertilization experiment in the western subarctic Pacific (SEEDS) by Atsushi Tsuda The response of N and Si to iron enrichment in the Northeast Pacific Ocean: Results from SERIES by David Timothy, C.S. Wong, Yukihiro Nojiri, Frank A. Whitney, W. Keith Johnson and Janet Barwell-Clarke 3.2 Biological and Physiological Responses (pdf, 0.2 Mb) Zooplankton responses during SEEDS by Hiroaki Saito Phytoplankton community response to iron and temperature gradient in the NW and NE subarctic Pacific Ocean by Isao Kudo, Yoshifumi Noiri, Jun Nishioka, Hiroshi Kiyosawa and Atsushi Tsuda SERIES: Copepod grazing on diatoms by Frank A. Whitney, Moira Galbraith, Janet Barwell-Clarke and Akash Sastri The Southern Ocean Iron Enrichment Experiment: The nitrogen uptake response by William P. Cochlan and Raphael M. Kudela 3.3 Biogeochemical Responses (pdf, 0.5 Mb) What have we learned regarding iron biogeochemistry from iron enrichment experiments? by Jun Nishioka, Shigenobu Takeda and W. Keith Johnson Iron dynamics and temporal changes of iron speciation in SERIES by W. Keith Johnson, C.S. Wong, Nes Sutherland and Jun Nishioka Dissolved organic matter dynamics during SEEDS and SERIES experiments by Takeshi Yoshimura and Hiroshi Ogawa Formation of transparent exopolymer particles during the in-situ iron enrichment experiment in the western subarctic Pacific (SEEDS) by Shigenobu Takeda, Neelam Ramaiah, Ken Furuya and Takeshi Yoshimura Atmospheric measurement by Mitsuo Uematsu 3.4 Prediction from Models (pdf, 0.3 Mb) Modelling iron limitation in the North Pacific by Kenneth L. Denman and M. Angelica Peña A proposed model of the SERIES iron fertilization patch by Debby Ianson, Christoph Voelker and Kenneth L. Denman 4. LIST OF PARTICIPANTS FOR THE 2004 WORKSHOP (pdf, < 0.1 Mb) APPENDIX 1 Report of the 2000 Planning Workshop on Designing the Iron Fertilization Experiment in the Subarctic Pacific (pdf, 1 Mb) APPENDIX 2 Terms of Reference for the Advisory Panel on Iron fertilization experiment in the subarctic Pacific Ocean (pdf, < 0.1 Mb) APPENDIX 3 Historical List of Advisory Panel Members on Iron fertilization experiment in the subarctic Pacific Ocean (pdf, < 0.1 Mb) APPENDIX 4 IFEP-AP Annual Reports (pdf, 0.1 Mb) APPENDIX 5 PICES Press Articles (pdf, 0.6 Mb) (194 page document)

Pratica dos inventarios, partilhas e contas : primeira parte dos juizos divisorios : obra necessaria aos que começão no auditorio forense falicitando o uso da materia divisoria nas diversas questões de cabeça de casal, collações, licitações, bemfeitorias de prazos e morgados, forma da repartição, lesões, emenda de partilhas, sonegados, recursos na 1. e 2. instancia, formaes, ou cartas de partilhas, testamentos, posses, heranças jacentes, contas no juizo dos orphãos e provedorias com uma tabella das acções civeis e outra dos processos criminaes classe dos delictos (sic)

Additamento a pratica dos inventarios, partilhas e contas dos juizos divisorios de Alberto Carlos de Menezes, em que se apontão as differenças entre a legislação anterior a 1833, e a que actualmente regula esta materia / por Joaquim Raphael do Valle.

Soundings: the Newsletter of the Monterey Bay Chapter of the American Cetacean Society. 2005

(PDF contains 89 pages.)

Poezias de Pedro de Andrade Caminha, mandadas publicar pela Academia Real das Sciencias de Lisboa

Pedro de Andrade Caminha, poeta português, nasceu no Porto em 1520 e morreu em Vila Viçosa em 1589. Apenas em 1791 a Academia das Ciencias reuniu e mandou publicar suas poesias num total de quatro centos e cinquenta e quatro trabalhos. Segundo Inocêncio, ‘a coleção foi feita sobre os manuscritos que separadamente existiam, um na livraria do Convento da Graça de Lisboa, contendo as éclogas, epístolas, odes, epitáfios e algumas elegias, etc.; outro na do Duque de Cadaval, que continha os epitalâmios, epigramas, outras elegias, e duas epístolas; e aproveitaram-se também alguns sonetos e outros pequenos poemas, que andavam sendo impressos desde o tempo do autor, mas incorporados em obras de diversos autores. Deste modo se completou esta primeira e única edição, dirigida pelos cuidados do então secretário da Academia José Correia da Serra, de quem é o prólogo que a precede’

Lei de acesso à informação (LAI) : investigação da transparência nos três Poderes

A Lei 12.527/2011, Lei de Acesso à Informação (LAI), representa um avanço na disponibilização da informação pública pelos órgãos governamentais, de maneira a promover a transparência das ações e a consolidação da cidadania. Este artigo tem como foco a observância do atendimento dos preceitos de transparência ativa pelos sítios oficiais de órgãos públicos federais dos Poderes Executivo, Legislativo e Judiciário.