The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Many machine learning tasks have been shown solvable with impressive levels of success given large amounts of training data and computational power. For the problems which lack data sufficient to achieve high performance, methods for transfer learning can be applied. These refer to performing the new task while having prior knowledge of the nature of the data, gained by first performing a different...
Deep neural networks have been widely applied in the field of environmental sound classification. However, due to the scarcity of carefully labeled data, their training process suffers from over-fitting. Data augmentation is a technique that alleviates this issue. It augments the training set with synthetic data that are created by modifying some parameters of the real data. However, not all kinds...
We propose a method for optimizing an acoustic feature extractor for anomalous sound detection (ASD). Most ASD systems adopt outlier-detection techniques because it is difficult to collect a massive amount of anomalous sound data. To improve the performance of such outlier-detection-based ASD, it is essential to extract a set of efficient acoustic features that is suitable for identifying anomalous...
This paper investigates the use of Dirichlet process hidden Markov model (DPHMM) tokenizer for the template matching based query-by-example spoken term detection (QbE-STD) task. DPHMM can be obtained following an unsupervised iterative procedure without any training transcriptions. The STD performance of the DPHMM tokenizer is evaluated on TIMIT Corpus. We construct three kinds of DPHMM based QbE-STD...
There are several challenges while building Automatic Speech Recognition (ASR) system for low resource languages such as Indic languages. One problem is the access to large amounts of training data required to build Acoustic Models (AM) from scratch. In the context of Indian English, another challenge encountered is code-mixing as many Indian speakers are multilingual and exhibit code-mixing in their...
With the completion of the IARPA Babel program, it is possible to systematically analyze the performance of speech recognition systems across a wide variety of languages. We select 16 languages from the dataset and compare performance using a deep neural network-based acoustic model. The focus is on keyword spotting using the actual term-weighted value (ATWV) metric. We demonstrate that ATWV is keyword...
In this paper we describe the 2016 BBN conversational telephone speech keyword spotting system; the culmination of four years of research and development under the IARPA Babel program. The system was constructed in response to the NIST Open Keyword Search (OpenKWS) evaluation of 2016. We present our technological breakthroughs in building top-performing keyword spotting processing systems for new...
Conventional deep neural networks (DNN) for speech acoustic modeling rely on Gaussian mixture models (GMM) and hidden Markov model (HMM) to obtain binary class labels as the targets for DNN training. Subword classes in speech recognition systems correspond to context-dependent tied states or senones. The present work addresses some limitations of GMM-HMM senone alignments for DNN training. We hypothesize...
The environmental robustness of DNN-based acoustic models can be significantly improved by using multi-condition training data. However, as data collection is a costly proposition, simulation of the desired conditions is a frequently adopted strategy. In this paper we detail a data augmentation approach for far-field ASR. We examine the impact of using simulated room impulse responses (RIRs), as real...
In many event detection applications, training data may contain tags with multiple, simultaneous events. This is particularly likely when the definition of “event” is broad and includes events that can persist for an extended period of time. Decomposing a mixed signal into signals corresponding to individual events is non-trivial. In this paper, we propose a non-negative matrix factorization (NMF)...
When using connectionist temporal classification (CTC) based acoustic models (AMs) for large vocabulary continuous speech recognition (LVCSR), most previous studies have used a naive interpolation of the CTC-AM score and an additional language model score, although there is no theoretical justification for such an approach. On the other hand, we recently proposed a theoretically more sound decoding...
In this work we explore data-augmentation techniques for the task of improving the performance of a supervised recurrent-neural-network classifier tasked with predicting prosodic-boundary and pitch-accent labels. The technique is based on applying voice transformations to the training data that modify the pitch baseline and range, as well as the vocal-tract and vocal-source characteristics of the...
DNN based acoustic models require a large amount of training data. Parametric data augmentation techniques such as adding noise, reverberation, or changing the speech rate, are often employed to boost the dataset size and the ASR performance. The choice of augmentation techniques and the associated parameters has been handled heuristically so far. In this work we propose an algorithm to automatically...
Batch normalization is a standard technique for training deep neural networks. In batch normalization, the input of each hidden layer is first mean-variance normalized and then linearly transformed before applying non-linear activation functions. We propose a novel unsupervised speaker adaptation technique for batch normalized acoustic models. The key idea is to adjust the linear transformations previously...
Multilingual (ML) representations play a key role in building speech recognition systems for low resource languages. The IARPA sponsored BABEL program focuses on building speech recognition (ASR) and keyword search (KWS) systems in over 24 languages with limited training data. The most common mechanism to derive ML representations in the BABEL program has been with the use of a two-stage network,...
This paper describes a new technique to automatically obtain large high-quality training speech corpora for acoustic modeling. Traditional approaches select utterances based on confidence thresholds and other heuristics. We propose instead to use an ensemble approach: we transcribe each utterance using several recognizers, and only keep those on which they agree. The recognizers we use are trained...
Deep neural network (DNN) acoustic models can be adapted to under-resourced languages by transferring the hidden layers. An analogous transfer problem is popular as few-shot learning to recognise scantily seen objects based on their meaningful attributes. In similar way, this paper proposes a principled way to represent the hidden layers of DNN in terms of attributes shared across languages. The diverse...
Acoustic model performance typically decreases when evaluated on a dialectal variation of the same language that was not used during training. Similarly, models simultaneously trained on a group of dialects tend to underperform dialect-specific models. In this paper, we report on our efforts towards building a unified acoustic model that can serve a multi-dialectal language. Two techniques are presented:...
Training very deep neural networks is very difficult because of gradient degradation. However, the incomparable expressiveness of the many deep layers is highly desirable at testing time and usually leads to better performance. Recently, training techniques such as residual networks that enable us to train very deep networks have proved to be a great success. In this paper, we studied the application...
Mismatched crowdsourcing is a technique to derive speech transcriptions using crowd-workers unfamiliar with the language being spoken. This technique is especially useful for under-resourced languages since it is hard to hire native transcribers. In this paper, we demonstrate that using mismatched transcription for adaptation improves performance of speech recognition under limited matched training...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.