Cross-lingual word embeddings can be applied to several natural language processing applications across multiple languages. Unlike prior works that use word embeddings based on the Euclidean space, this short paper presents a simple and effective cross-lingual Word2Vec model that adapts to the Poincar\'e ball model of hyperbolic space to learn unsupervised cross-lingual word representations from a German-English parallel corpus. It has been shown that hyperbolic embeddings can capture and preserve hierarchical relationships. We evaluate the model on both hypernymy and analogy tasks. The proposed model achieves comparable performance with the vanilla Word2Vec model on the cross-lingual analogy task, the hypernymy task shows that the cross-lingual Poincar\'e Word2Vec model can capture latent hierarchical structure from free text across languages, which are absent from the Euclidean-based Word2Vec representations. Our results show that by preserving the latent hierarchical information, hyperbolic spaces can offer better representations for cross-lingual embeddings.
Despite the rapid advance of automatic speech recognition (ASR) technologies, accurate recognition of cocktail party speech characterised by the interference from overlapping speakers, background noise and room reverberation remains a highly challenging task to date. Motivated by the invariance of visual modality to acoustic signal corruption, audio-visual speech enhancement techniques have been developed, although predominantly targeting overlapping speech separation and recognition tasks. In this paper, an audio-visual multi-channel speech separation, dereverberation and recognition approach featuring a full incorporation of visual information into all three stages of the system is proposed. The advantage of the additional visual modality over using audio only is demonstrated on two neural dereverberation approaches based on DNN-WPE and spectral mapping respectively. The learning cost function mismatch between the separation and dereverberation models and their integration with the back-end recognition system is minimised using fine-tuning on the MSE and LF-MMI criteria. Experiments conducted on the LRS2 dataset suggest that the proposed audio-visual multi-channel speech separation, dereverberation and recognition system outperforms the baseline audio-visual multi-channel speech separation and recognition system containing no dereverberation module by a statistically significant word error rate (WER) reduction of 2.06% absolute (8.77% relative).
Previous works on expressive speech synthesis mainly focus on current sentence. The context in adjacent sentences is neglected, resulting in inflexible speaking style for the same text, which lacks speech variations. In this paper, we propose a hierarchical framework to model speaking style from context. A hierarchical context encoder is proposed to explore a wider range of contextual information considering structural relationship in context, including inter-phrase and inter-sentence relations. Moreover, to encourage this encoder to learn style representation better, we introduce a novel training strategy with knowledge distillation, which provides the target for encoder training. Both objective and subjective evaluations on a Mandarin lecture dataset demonstrate that the proposed method can significantly improve the naturalness and expressiveness of the synthesized speech.
Previous works on expressive speech synthesis focus on modelling the mono-scale style embedding from the current sentence or context, but the multi-scale nature of speaking style in human speech is neglected. In this paper, we propose a multi-scale speaking style modelling method to capture and predict multi-scale speaking style for improving the naturalness and expressiveness of synthetic speech. A multi-scale extractor is proposed to extract speaking style embeddings at three different levels from the ground-truth speech, and explicitly guide the training of a multi-scale style predictor based on hierarchical context information. Both objective and subjective evaluations on a Mandarin audiobooks dataset demonstrate that our proposed method can significantly improve the naturalness and expressiveness of the synthesized speech.
Automatic recognition of dysarthric and elderly speech highly challenging tasks to date. Speaker-level heterogeneity attributed to accent or gender commonly found in normal speech, when aggregated with age and speech impairment severity, create large diversity among speakers. Speaker adaptation techniques play a crucial role in personalization of ASR systems for such users. Their mobility issues limit the amount of speaker-level data available for model based adaptation. To this end, this paper investigates two novel forms of feature based on-the-fly rapid speaker adaptation approaches. The first is based on speaker-level variance regularized spectral basis embedding (SBEVR) features, while the other uses on-the-fly learning hidden unit contributions (LHUC) transforms conditioned on speaker-level spectral features. Experiments conducted on the UASpeech dysarthric and DimentiaBank Pitt elderly speech datasets suggest the proposed SBEVR features based adaptation statistically significantly outperform both the baseline on-the-fly i-Vector adapted hybrid TDNN/DNN systems by up to 2.48% absolute (7.92% relative) reduction in word error rate (WER), and offline batch mode model based LHUC adaptation using all speaker-level data by 0.78% absolute (2.41% relative) in WER reduction.
Zero-shot speaker adaptation aims to clone an unseen speaker's voice without any adaptation time and parameters. Previous researches usually use a speaker encoder to extract a global fixed speaker embedding from reference speech, and several attempts have tried variable-length speaker embedding. However, they neglect to transfer the personal pronunciation characteristics related to phoneme content, leading to poor speaker similarity in terms of detailed speaking styles and pronunciation habits. To improve the ability of the speaker encoder to model personal pronunciation characteristics, we propose content-dependent fine-grained speaker embedding for zero-shot speaker adaptation. The corresponding local content embeddings and speaker embeddings are extracted from a reference speech, respectively. Instead of modeling the temporal relations, a reference attention module is introduced to model the content relevance between the reference speech and the input text, and to generate the fine-grained speaker embedding for each phoneme encoder output. The experimental results show that our proposed method can improve speaker similarity of synthesized speeches, especially for unseen speakers.
Text normalization, defined as a procedure transforming non standard words to spoken-form words, is crucial to the intelligibility of synthesized speech in text-to-speech system. Rule-based methods without considering context can not eliminate ambiguation, whereas sequence-to-sequence neural network based methods suffer from the unexpected and uninterpretable errors problem. Recently proposed hybrid system treats rule-based model and neural model as two cascaded sub-modules, where limited interaction capability makes neural network model cannot fully utilize expert knowledge contained in the rules. Inspired by Flat-LAttice Transformer (FLAT), we propose an end-to-end Chinese text normalization model, which accepts Chinese characters as direct input and integrates expert knowledge contained in rules into the neural network, both contribute to the superior performance of proposed model for the text normalization task. We also release a first publicly accessible largescale dataset for Chinese text normalization. Our proposed model has achieved excellent results on this dataset.
Deep neural networks have brought significant advancements to speech emotion recognition (SER). However, the architecture design in SER is mainly based on expert knowledge and empirical (trial-and-error) evaluations, which is time-consuming and resource intensive. In this paper, we propose to apply neural architecture search (NAS) techniques to automatically configure the SER models. To accelerate the candidate architecture optimization, we propose a uniform path dropout strategy to encourage all candidate architecture operations to be equally optimized. Experimental results of two different neural structures on IEMOCAP show that NAS can improve SER performance (54.89\% to 56.28\%) while maintaining model parameter sizes. The proposed dropout strategy also shows superiority over the previous approaches.
The accuracy of prosodic structure prediction is crucial to the naturalness of synthesized speech in Mandarin text-to-speech system, but now is limited by widely-used sequence-to-sequence framework and error accumulation from previous word segmentation results. In this paper, we propose a span-based Mandarin prosodic structure prediction model to obtain an optimal prosodic structure tree, which can be converted to corresponding prosodic label sequence. Instead of the prerequisite for word segmentation, rich linguistic features are provided by Chinese character-level BERT and sent to encoder with self-attention architecture. On top of this, span representation and label scoring are used to describe all possible prosodic structure trees, of which each tree has its corresponding score. To find the optimal tree with the highest score for a given sentence, a bottom-up CKY-style algorithm is further used. The proposed method can predict prosodic labels of different levels at the same time and accomplish the process directly from Chinese characters in an end-to-end manner. Experiment results on two real-world datasets demonstrate the excellent performance of our span-based method over all sequence-to-sequence baseline approaches.
Although deep learning and end-to-end models have been widely used and shown superiority in automatic speech recognition (ASR) and text-to-speech (TTS) synthesis, state-of-the-art forced alignment (FA) models are still based on hidden Markov model (HMM). HMM has limited view of contextual information and is developed with long pipelines, leading to error accumulation and unsatisfactory performance. Inspired by the capability of attention mechanism in capturing long term contextual information and learning alignments in ASR and TTS, we propose a neural network based end-to-end forced aligner called NeuFA, in which a novel bidirectional attention mechanism plays an essential role. NeuFA integrates the alignment learning of both ASR and TTS tasks in a unified framework by learning bidirectional alignment information from a shared attention matrix in the proposed bidirectional attention mechanism. Alignments are extracted from the learnt attention weights and optimized by the ASR, TTS and FA tasks in a multi-task learning manner. Experimental results demonstrate the effectiveness of our proposed model, with mean absolute error on test set drops from 25.8 ms to 23.7 ms at word level, and from 17.0 ms to 15.7 ms at phoneme level compared with state-of-the-art HMM based model.