In hybrid HMM based speech recognition, LSTM language models have been widely applied and achieved large improvements. The theoretical capability of modeling any unlimited context suggests that no recombination should be applied in decoding. This motivates to reconsider full summation over the HMM-state sequences instead of Viterbi approximation in decoding. We explore the potential gain from more accurate probabilities in terms of decision making and apply the full-sum decoding with a modified prefix-tree search framework. The proposed full-sum decoding is evaluated on both Switchboard and Librispeech corpora. Different models using CE and sMBR training criteria are used. Additionally, both MAP and confusion network decoding as approximated variants of general Bayes decision rule are evaluated. Consistent improvements over strong baselines are achieved in almost all cases without extra cost. We also discuss tuning effort, efficiency and some limitations of full-sum decoding.
Acoustic models based on long short-term memory recurrent neural networks (LSTM-RNNs) were applied to statistical parametric speech synthesis (SPSS) and showed significant improvements in naturalness and latency over those based on hidden Markov models (HMMs). This paper describes further optimizations of LSTM-RNN-based SPSS for deployment on mobile devices; weight quantization, multi-frame inference, and robust inference using an ε-contaminated Gaussian loss function. Experimental results in subjective listening tests show that these optimizations can make LSTM-RNN-based SPSS comparable to HM…
University of Erlangen-Nuremberg（埃尔朗根-纽伦堡大学）；INESC-ID-Lisboa（里斯本INESC-ID研究所）
In this paper, we advance a recently-proposed uncertainty decoding scheme for DNN-HMM (deep neural network - hidden Markov model) hybrid systems. This numerical sampling concept averages DNN outputs produced by a finite set of feature samples (drawn from a probabilistic distortion model) to approximate the posterior likelihoods of the context-dependent HMM states. As main innovation, we propose a weighted DNN-output averaging based on a minimum classification error criterion and apply it to a probabilistic distortion model for spatial diffuseness features. The experimental evaluation is perfor…
Swiss AI Lab, IDSIA（瑞士人工智能实验室（IDSIA））；USI（瑞士意大利语区大学）；SUPSI（瑞士意大利语区应用科学大学）
We present a complete training pipeline to build a state-of-the-art hybrid HMM-based ASR system on the 2nd release of the TED-LIUM corpus. Data augmentation using SpecAugment is successfully applied to improve performance on top of our best SAT model using i-vectors. By investigating the effect of different maskings, we achieve improvements from SpecAugment on hybrid HMM models without increasing model size and training time. A subsequent sMBR training is applied to fine-tune the final acoustic model, and both LSTM and Transformer language models are trained and evaluated. Our best system achi…
In recent years, the standard hybrid DNN-HMM speech recognizers are outperformed by the end-to-end speech recognition systems. One of the very promising approaches is the grapheme Wav2Vec 2.0 model, which uses the self-supervised pretraining approach combined with transfer learning of the fine-tuned speech recognizer. Since it lacks the pronunciation vocabulary and language model, the approach is suitable for tasks where obtaining such models is not easy or almost impossible. In this paper, we use the Wav2Vec speech recognizer in the task of spoken term detection over a large set of spoken doc…
We propose an information theoretic framework for quantitative assessment of acoustic modeling for hidden Markov model (HMM) based automatic speech recognition (ASR). Acoustic modeling yields the probabilities of HMM sub-word states for a short temporal window of speech acoustic features. We cast ASR as a communication channel where the input sub-word probabilities convey the information about the output HMM state sequence. The quality of the acoustic model is thus quantified in terms of the information transmitted through this channel. The process of inferring the most likely HMM state sequen…
We present state-of-the-art automatic speech recognition (ASR) systems employing a standard hybrid DNN/HMM architecture compared to an attention-based encoder-decoder design for the LibriSpeech task. Detailed descriptions of the system development, including model design, pretraining schemes, training schedules, and optimization approaches are provided for both system architectures. Both hybrid DNN/HMM and attention-based systems employ bi-directional LSTMs for acoustic modeling/encoding. For language modeling, we employ both LSTM and Transformer based architectures. All our systems are built …
We investigate the impact of aggressive low-precision representations of weights and activations in two families of large LSTM-based architectures for Automatic Speech Recognition (ASR): hybrid Deep Bidirectional LSTM - Hidden Markov Models (DBLSTM-HMMs) and Recurrent Neural Network - Transducers (RNN-Ts). Using a 4-bit integer representation, a naïve quantization approach applied to the LSTM portion of these models results in significant Word Error Rate (WER) degradation. On the other hand, we show that minimal accuracy loss is achievable with an appropriate choice of quantizers and initializ…
This paper presents a novel deep learning architecture for acoustic model in the context of Automatic Speech Recognition (ASR), termed as MixNet. Besides the conventional layers, such as fully connected layers in DNN-HMM and memory cells in LSTM-HMM, the model uses two additional layers based on Mixture of Experts (MoE). The first MoE layer operating at the input is based on pre-defined broad phonetic classes and the second layer operating at the penultimate layer is based on automatically learned acoustic classes. In natural speech, overlap in distribution across different acoustic classes is…
Brain Machine Interface Lab（脑机接口实验室）；The University of Texas at Austin（得克萨斯大学奥斯汀分校）
In this paper we first demonstrate continuous noisy speech recognition using electroencephalography (EEG) signals on English vocabulary using different types of state of the art end-to-end automatic speech recognition (ASR) models, we further provide results obtained using EEG data recorded under different experimental conditions. We finally demonstrate decoding of speech spectrum from EEG signals using a long short term memory (LSTM) based regression model and Generative Adversarial Network (GAN) based model. Our results demonstrate the feasibility of using EEG signals for continuous noisy sp…
Considering audio and image data as having quantum nature (data are represented by density matrices), we achieved better results on training architectures such as 3-layer stacked LSTM and HMM by mixing training samples using superposition augmentation and compared with plain default training and mix-up augmentation. This augmentation technique originates from the mix-up approach but provides more solid theoretical reasoning based on quantum properties. We achieved 3% improvement (from 68% to 71%) by using 38% lesser number of training samples in Russian audio-digits recognition task and 7,16% …
This paper proposes a simple and effective approach for automatic recognition of Cued Speech (CS), a visual communication tool that helps people with hearing impairment to understand spoken language with the help of hand gestures that can uniquely identify the uttered phonemes in complement to lipreading. The proposed approach is based on a pre-trained hand and lips tracker used for visual feature extraction and a phonetic decoder based on a multistream recurrent neural network trained with connectionist temporal classification loss and combined with a pronunciation lexicon. The proposed syste…
This work concentrates on reducing the RTF and word error rate of a hybrid HMM-DNN. Our baseline system uses an architecture with TDNN and LSTM layers. We find this architecture particularly useful for lightly reverberated environments. However, these models tend to demand more computation than is desirable. In this work, we explore alternate architectures employing singular value decomposition (SVD) is applied to the TDNN layers to reduce the RTF, as well as to the affine transforms of every LSTM cell. We compare this approach with specifying bottleneck layers similar to those introduced by S…
Human can recognize speech, as well as the peculiar accent of the speech simultaneously. However, present state-of-the-art ASR system can rarely do that. In this paper, we propose a multilingual approach to recognizing English speech, and related accent that speaker conveys using DNN-HMM framework. Specifically, we assume different accents of English as different languages. We then merge them together and train a multilingual ASR system. During decoding, we conduct two experiments. One is a monolingual ASR-based decoding, with the accent information embedded at phone level, realizing word-base…
In automatic speech recognition (ASR) research, discriminative criteria have achieved superior performance in DNN-HMM systems. Given this success, the adoption of discriminative criteria is promising to boost the performance of end-to-end (E2E) ASR systems. With this motivation, previous works have introduced the minimum Bayesian risk (MBR, one of the discriminative criteria) into E2E ASR systems. However, the effectiveness and efficiency of the MBR-based methods are compromised: the MBR criterion is only used in system training, which creates a mismatch between training and decoding; the on-t…
We discuss the problem of echographic transcription in autoregressive sequence-to-sequence attentional architectures for automatic speech recognition, where a model produces very long sequences of repetitive outputs when presented with out-of-domain utterances. We decode audio from the British National Corpus with an attentional encoder-decoder model trained solely on the LibriSpeech corpus. We observe that there are many 5-second recordings that produce more than 500 characters of decoding output (i.e. more than 100 characters per second). A frame-synchronous hybrid (DNN-HMM) model trained on…
End-to-end (E2E) systems have played a more and more important role in automatic speech recognition (ASR) and achieved great performance. However, E2E systems recognize output word sequences directly with the input acoustic feature, which can only be trained on limited acoustic data. The extra text data is widely used to improve the results of traditional artificial neural network-hidden Markov model (ANN-HMM) hybrid systems. The involving of extra text data to standard E2E ASR systems may break the E2E property during decoding. In this paper, a novel modular E2E ASR system is proposed. The mo…
In our previous work we demonstrated that a single headed attention encoder-decoder model is able to reach state-of-the-art results in conversational speech recognition. In this paper, we further improve the results for both Switchboard 300 and 2000. Through use of an improved optimizer, speaker vector embeddings, and alternative speech representations we reduce the recognition errors of our LSTM system on Switchboard-300 by 4% relative. Compensation of the decoder model with the probability ratio approach allows more efficient integration of an external language model, and we report 5.9% and …
We introduce "Unspeech" embeddings, which are based on unsupervised learning of context feature representations for spoken language. The embeddings were trained on up to 9500 hours of crawled English speech data without transcriptions or speaker information, by using a straightforward learning objective based on context and non-context discrimination with negative sampling. We use a Siamese convolutional neural network architecture to train Unspeech embeddings and evaluate them on speaker comparison, utterance clustering and as a context feature in TDNN-HMM acoustic models trained on TED-LIUM,…
Beam search, which is the dominant ASR decoding algorithm for end-to-end models, generates tree-structured hypotheses. However, recent studies have shown that decoding with hypothesis merging can achieve a more efficient search with comparable or better performance. But, the full context in recurrent networks is not compatible with hypothesis merging. We propose to use vector-quantized long short-term memory units (VQ-LSTM) in the prediction network of RNN transducers. By training the discrete representation jointly with the ASR network, hypotheses can be actively merged for lattice generation…
LSTM based language models are an important part of modern LVCSR systems as they significantly improve performance over traditional backoff language models. Incorporating them efficiently into decoding has been notoriously difficult. In this paper we present an approach based on a combination of one-pass decoding and lattice rescoring. We perform decoding with the LSTM-LM in the first pass but recombine hypothesis that share the last two words, afterwards we rescore the resulting lattice. We run our systems on GPGPU equipped machines and are able to produce competitive results on the Hub5'00 a…
This article describes a density ratio approach to integrating external Language Models (LMs) into end-to-end models for Automatic Speech Recognition (ASR). Applied to a Recurrent Neural Network Transducer (RNN-T) ASR model trained on a given domain, a matched in-domain RNN-LM, and a target domain RNN-LM, the proposed method uses Bayes' Rule to define RNN-T posteriors for the target domain, in a manner directly analogous to the classic hybrid model for ASR based on Deep Neural Networks (DNNs) or LSTMs in the Hidden Markov Model (HMM) framework (Bourlard & Morgan, 1994). The proposed approach i…
Hybrid automatic speech recognition (ASR) models are typically sequentially trained with CTC or LF-MMI criteria. However, they have vastly different legacies and are usually implemented in different frameworks. In this paper, by decoupling the concepts of modeling units and label topologies and building proper numerator/denominator graphs accordingly, we establish a generalized framework for hybrid acoustic modeling (AM). In this framework, we show that LF-MMI is a powerful training criterion applicable to both limited-context and full-context models, for wordpiece/mono-char/bi-char/chenone un…
Decoding the speech signal that a person is listening to from the human brain via electroencephalography (EEG) can help us understand how our auditory system works. Linear models have been used to reconstruct the EEG from speech or vice versa. Recently, Artificial Neural Networks (ANNs) such as Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) based architectures have outperformed linear models in modeling the relation between EEG and speech. Before attempting to use these models in real-world applications such as hearing tests or (second) language comprehension assessment w…
Human brain performs remarkably well in segregating a particular speaker from interfering ones in a multi-speaker scenario. It has been recently shown that we can quantitatively evaluate the segregation capability by modelling the relationship between the speech signals present in an auditory scene and the cortical signals of the listener measured using electroencephalography (EEG). This has opened up avenues to integrate neuro-feedback into hearing aids whereby the device can infer user's attention and enhance the attended speaker. Commonly used algorithms to infer the auditory attention are …
In this work, we conduct an extensive comparison of various approaches to speech based emotion recognition systems. The analyses were carried out on audio recordings from Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS). After pre-processing the raw audio files, features such as Log-Mel Spectrogram, Mel-Frequency Cepstral Coefficients (MFCCs), pitch and energy were considered. The significance of these features for emotion classification was compared by applying methods such as Long Short Term Memory (LSTM), Convolutional Neural Networks (CNNs), Hidden Markov Models (HMMs) …
Modeling the relationship between natural speech and a recorded electroencephalogram (EEG) helps us understand how the brain processes speech and has various applications in neuroscience and brain-computer interfaces. In this context, so far mainly linear models have been used. However, the decoding performance of the linear model is limited due to the complex and highly non-linear nature of the auditory processing in the human brain. We present a novel Long Short-Term Memory (LSTM)-based architecture as a non-linear model for the classification problem of whether a given pair of (EEG, speech …
