Search

This work addresses the use of deep neural networks (DNNs) in automatic language identification (LID) focused on short test utterances. Motivated by their recent success in acoustic modelling for speech recognition, we adapt DNNs to the problem of identifying the language in a given utterance from the short-term acoustic features. We show how DNNs are particularly suitable to perform LID in real-time applications, due to their capacity to emit a language identification posterior at each new frame of the test utterance. We then analyse different aspects of the system, such as the amount of required training data, the number of hidden layers, the relevance of contextual information and the effect of the test utterance duration. Finally, we propose several methods to combine frame-by-frame posteriors. Experiments are conducted on two different datasets: the public NIST Language Recognition Evaluation 2009 (3 s task) and a much larger corpus (of 5 million utterances) known as Google 5M LID, obtained from different Google Services. Reported results show relative improvements of DNNs versus the i-vector system of 40% in LRE09 3 second task and 76% in Google 5M LID., (Copyright © 2014 Elsevier Ltd. All rights reserved.)

Long Short Term Memory (LSTM) Recurrent Neural Networks (RNNs) have recently outperformed other state-of-the-art approaches, such as i-vector and Deep Neural Networks (DNNs), in automatic Language Identification (LID), particularly when dealing with very short utterances (∼3s). In this contribution we present an open-source, end-to-end, LSTM RNN system running on limited computational resources (a single GPU) that outperforms a reference i-vector system on a subset of the NIST Language Recognition Evaluation (8 target languages, 3s task) by up to a 26%. This result is in line with previously published research using proprietary LSTM implementations and huge computational resources, which made these former results hardly reproducible. Further, we extend those previous experiments modeling unseen languages (out of set, OOS, modeling), which is crucial in real applications. Results show that a LSTM RNN with OOS modeling is able to detect these languages and generalizes robustly to unseen OOS languages. Finally, we also analyze the effect of even more limited test data (from 2.25s to 0.1s) proving that with as little as 0.5s an accuracy of over 50% can be achieved. [ABSTRACT FROM AUTHOR]

Automatic speech recognition (ASR) systems are used daily by millions of people worldwide to dictate messages, control devices, initiate searches or to facilitate data input in small devices. The user experience in these scenarios depends on the quality of the speech transcriptions and on the responsiveness of the system. For multilingual users, a further obstacle to natural interaction is the monolingual character of many ASR systems, in which users are constrained to a single preset language. In this work, we present an end-to-end multi-language ASR architecture, developed and deployed at Google, that allows users to select arbitrary combinations of spoken languages. We leverage recent advances in language identification and a novel method of real-time language selection to achieve similar recognition accuracy and nearly-identical latency characteristics as a monolingual system. [ABSTRACT FROM PUBLISHER]

This paper presents the systems submitted by the ATVS Biometric Recognition Group to the 2009 Language Recognition Evaluation (LRE'09), organized by NIST. New challenges included in this LRE edition can be summarized by three main differences with respect to past evaluations. First, the number of languages to be recognized expanded to 23 languages from 14 in 2007, and 7 in 2005. Second, the data variability has been increased by including telephone speech excerpts extracted from Voice of America (VOA) radio broadcasts through Internet in addition to conversational telephone speech (CTS). The third difference was the volume of data, involving in this evaluation up to 2 terabytes of speech data for development, which is an order of magnitude greater than past evaluations. LRE'09 thus required participants to develop robust systems able not only to successfully face the session variability problem but also to do it with reasonable computational resources. ATVS participation consisted of state-of-the-art acoustic and high-level systems focussing on these issues. Furthermore, the problem of finding a proper combination and calibration of the information obtained at different levels of the speech signal was widely explored in this submission. In this paper, two original contributions were developed. The first contribution was applying a session variability compensation scheme based on factor analysis (FA) within the statistics domain into a SVM-supervector (SVM-SV) approach. The second contribution was the employment of a novel back-end based on anchor models in order to fuse individual systems prior to one-versus-all calibration via logistic regression. Results both in development and evaluation corpora show the robustness and excellent performance of the submitted systems, exemplified by our system ranked second in the 30-second open-set condition, with remarkably scarce computational resources. [ABSTRACT FROM PUBLISHER]