Exploració per autor "Rodríguez Fonollosa, José Adrián"
Ara es mostren els items 41-60 de 128
-
Combining subword representations into word-level representations in the transformer architecture
Casas Manzanares, Noé; Ruiz Costa-Jussà, Marta; Rodríguez Fonollosa, José Adrián (Association for Computational Linguistics, 2020)
Comunicació de congrés
Accés obertIn Neural Machine Translation, using word-level tokens leads to degradation in translation quality. The dominant approaches use subword-level tokens, but this increases the length of the sequences and makes it difficult ... -
Conditional distribution variability measures for causality detection
Rodríguez Fonollosa, José Adrián (Springer, 2019)
Capítol de llibre
Accés restringit per política de l'editorialIn this paper we derive variability measures for the conditional probability distributions of a pair of random variables, and we study its application in the inference of causal-effect relationships. We also study the ... -
Correcting input noise in SMT as a char-based translation problem
Formiga Fanals, Lluís; Rodríguez Fonollosa, José Adrián (2012-10-31)
Report de recerca
Accés obertMisspelled words have a direct impact on the final quality obtained by Statistical Machine Translation (SMT) systems as the input becomes noisy and unpredictable. This paper presents some improvement strategies for translating ... -
Coupling hierarchical word reordering and decoding in phrase-based statistical machine translation
Dras, Mark; Khalilov, Maxim; Rodríguez Fonollosa, José Adrián (2009-06)
Comunicació de congrés
Accés obertIn this paper, we start with the existing idea of taking reordering rules automatically derived from syntactic representations, and applying them in a preprocessing step before translation to make the source sentence ... -
Cuantificación vectorial adaptativa de la voz
Rodríguez Fonollosa, José Adrián; Masgrau Gómez, Enrique José; Carbonell, Rafael M. (1989)
Text en actes de congrés
Accés obertVector quantization (VQ) is a simultaneous quantization of a sequence of samples or vector. This process allows to malee effective use of the interrelations among the different vector components and performance arbitrarily ... -
Cuantificación vectorial en codificación de voz por excitación multipulso
Moreno Bilbao, M. Asunción; Rodríguez Fonollosa, José Adrián (1991)
Text en actes de congrés
Accés obertIn this paper is made a comparison between sorne quantizers applied in a Multipulse Speech coder. Vector quantizer is compared against scalar quantization in the LPC parameter in the short predictor. Adaptive Multistage ... -
Dealing with input noise in statistical machine translation
Formiga Fanals, Lluís; Rodríguez Fonollosa, José Adrián (2012)
Comunicació de congrés
Accés obertMisspelled words have a direct impact on the final quality obtained by Statistical Machine Translation (SMT) systems as the input becomes noisy and unpredictable. This paper presents some improvement strategies for ... -
DeepVoice: tecnologías de aprendizaje profundo aplicadas al procesado de voz y audio
Ruiz Costa-Jussà, Marta; Rodríguez Fonollosa, José Adrián (2017-09-01)
Article
Accés obertThis project proposes the development of new deep learning methods for speech and audio processing, exploring new applications and continuing the initial work of the research team and the international community. Research ... -
DeepVoice: tecnologías de aprendizaje profundo aplicadas al procesado de voz y audio
Ruiz Costa-Jussà, Marta; Rodríguez Fonollosa, José Adrián (2017-09-22)
Article
Accés obertEste proyecto propone el desarrollo de nuevas arquitecturas para el procesado de la voz y el audio mediante métodos de aprendizaje profundo, explorando también nuevas aplicaciones y dando continuidad al trabajo inicial del ... -
Determining CPU and memory requirements for real-time speech recognition systems using the TMS320C3x/C4x
Batlle Mont, Eloi; Rodríguez Fonollosa, José Adrián (Texas Instruments (TI), 1996)
Text en actes de congrés
Accés restringit per política de l'editorialDeveloping a computer system using real-time speech recognition previously required a workstation using non-specialized CPUs. Limits to the system were imposed by the amount of memory and hardware required. The Texas ... -
End-to-end speech translation with pre-trained models and adapters: UPC at IWSLT 2021
Gallego Olsina, Gerard Ion; Tsiamas, Ioannis; Escolano Peinado, Carlos; Rodríguez Fonollosa, José Adrián; Ruiz Costa-Jussà, Marta (Association for Computational Linguistics, 2021)
Text en actes de congrés
Accés obertThis paper describes the submission to the IWSLT 2021 offline speech translation task by the UPC Machine Translation group. The task consists of building a system capable of translating English audio recordings extracted ... -
End-to-end speech translation with the transformer
Cross Vila, Laura; Escolano Peinado, Carlos; Rodríguez Fonollosa, José Adrián; Ruiz Costa-Jussà, Marta (Antonio Bonafonte, Jordi Luque and Francesc Alías Pujol, 2018)
Comunicació de congrés
Accés restringit per política de l'editorialSpeech Translation has been traditionally addressed with the concatenation of two tasks: Speech Recognition and Machine Translation. This approach has the main drawback that errors are concatenated. Recently, neural ... -
English-Latvian SMT: the challenge of translating into a free word order language
Khalilov, Maxim; Rodríguez Fonollosa, José Adrián; Skadina, Inguna; Bralitis, Edgar; Pretkalnina, Lauma (2010)
Comunicació de congrés
Accés obertThis paper presents a comparative study of two approaches to statistical machine translation (SMT) and their application to a task of English-to-Latvian translation, which is still an open research line in the field of ... -
Enhancing sequence-to-sequence modeling for RDF triples to natural text
Domingo Roig, Oriol; Bergés Lladó, David; Cantenys Sabà, Roser; Creus Castanyer, Roger; Rodríguez Fonollosa, José Adrián (Association for Computational Linguistics, 2020)
Text en actes de congrés
Accés obertEstablishes key guidelines on how, which and when Machine Translation (MT) techniques are worth applying to RDF-to-Text task. Not only do we apply and compare the most prominent MT architecture, the Transformer, but we ... -
Estimation of the modulation index of cpm signals using hos
Rodríguez Fonollosa, Javier; Rodríguez Fonollosa, José Adrián (Institute of Electrical and Electronics Engineers (IEEE), 1993)
Text en actes de congrés
Accés obertThree simple methods are proposed for the estimation of the modulation index of continuous phase modulated signals in noise. These methods employ the estimated autocorrelation and fourth-order cumulant sequences of the ... -
Feature decorrelation methods in speech recognition. A comparative study
Batlle Mont, Eloi; Nadeu Camprubí, Climent; Rodríguez Fonollosa, José Adrián (International Speech Communication Association (ISCA), 1998)
Text en actes de congrés
Accés obertIn this paper we study various decorrelation methods for the features used in speech recognition and we compare the performance of each one by running several tests with a speech database. First of all we study the ... -
Fir system identification using a linear combination of cumulants
Rodríguez Fonollosa, José Adrián; Vidal Manzano, José; Moreno Bilbao, M. Asunción (. IEEE INT. CONF. ON ACOUSTICS, SPEECH & SIGNAL PROC, 1992)
Text en actes de congrés
Accés obertA general linear approach to identifying the parameters of a moving average (MA) model from the statistics of the output is developed. It is shown that, under some constraints, the impulse response of the system can be ... -
First experiments on an HMM based double layer framework for automatic continuous speech recognition
Nogueiras Rodríguez, Albino; Casar López, Marta; Rodríguez Fonollosa, José Adrián; Caballero Galeote, Mónica (2006)
Comunicació de congrés
Accés restringit per política de l'editorialThe usual approach to automatic continuous speech recognition is what can be called the acoustic-phonetic modelling approach. In this approach, voice is considered to hold two different kinds of information acoustic and ... -
From bilingual to multilingual neural machine translation by incremental training
Escolano Peinado, Carlos; Ruiz Costa-Jussà, Marta; Rodríguez Fonollosa, José Adrián (Association for Computational Linguistics, 2019)
Comunicació de congrés
Accés obertMultilingual Neural Machine Translation approaches are based on the use of task specific models and the addition of one more language can only be done by retraining the whole system. In this work, we propose a new training ... -
From bilingual to multilingual neural-based machine translation by incremental training
Escolano Peinado, Carlos; Ruiz Costa-Jussà, Marta; Rodríguez Fonollosa, José Adrián (2020-08-02)
Article
Accés obertA common intermediate language representation in neural machine translation can be used to extend bilingual systems by incremental training. We propose a new architecture based on introducing an interlingual loss as an ...