Speech-driven visual speech synthesis involves mapping features extracted from acoustic speech to the corresponding lip animation controls for a face model. This mapping can take many forms, but a powerful approach is to use deep neural networks (DNNs). However, a limitation is the lack of synchronized audio, video, and depth data required to reliably train the DNNs, especially for speaker-independent models. In this paper, we investigate adapting an automatic speech recognition (ASR) acoustic model (AM) for the visual speech synthesis problem. We train the AM on ten thousand hours of audio-only data. The AM is then adapted to the visual speech synthesis domain using ninety hours of synchronized audio-visual speech. Using a subjective assessment test, we compared the performance of the AM-initialized DNN to one with a random initialization. The results show that viewers significantly prefer animations generated from the AM-initialized DNN than the ones generated using the randomly initialized model. We conclude that visual speech synthesis can significantly benefit from the powerful representation of speech in the ASR acoustic models.
Direct Speech-to-speech translation (S2ST) has drawn more and more attention recently. The task is very challenging due to data scarcity and complex speech-to-speech mapping. In this paper, we report our recent achievements in S2ST. Firstly, we build a S2ST Transformer baseline which outperforms the original Translatotron. Secondly, we utilize the external data by pseudo-labeling and obtain a new state-of-the-art result on the Fisher English-to-Spanish test set. Indeed, we exploit the pseudo data with a combination of popular techniques which are not trivial when applied to S2ST. Moreover, we evaluate our approach on both syntactically similar (Spanish-English) and distant (English-Chinese) language pairs. Our implementation is available at https://github.com/fengpeng-yue/speech-to-speech-translation.
This work presents our end-to-end (E2E) automatic speech recognition (ASR) model targetting at robust speech recognition, called Integraded speech Recognition with enhanced speech Input for Self-supervised learning representation (IRIS). Compared with conventional E2E ASR models, the proposed E2E model integrates two important modules including a speech enhancement (SE) module and a self-supervised learning representation (SSLR) module. The SE module enhances the noisy speech. Then the SSLR module extracts features from enhanced speech to be used for speech recognition (ASR). To train the proposed model, we establish an efficient learning scheme. Evaluation results on the monaural CHiME-4 task show that the IRIS model achieves the best performance reported in the literature for the single-channel CHiME-4 benchmark (2.0% for the real development and 3.9% for the real test) thanks to the powerful pre-trained SSLR module and the fine-tuned SE module.
Collecting speech data is an important step in training speech recognition systems and other speech-based machine learning models. However, the issue of privacy protection is an increasing concern that must be addressed. The current study investigates the use of voice conversion as a method for anonymizing voices. In particular, we train several voice conversion models using self-supervised speech representations including Wav2Vec2.0, Hubert and UniSpeech. Converted voices retain a low word error rate within 1% of the original voice. Equal error rate increases from 1.52% to 46.24% on the LibriSpeech test set and from 3.75% to 45.84% on speakers from the VCTK corpus which signifies degraded performance on speaker verification. Lastly, we conduct experiments on dysarthric speech data to show that speech features relevant to articulation, prosody, phonation and phonology can be extracted from anonymized voices for discriminating between healthy and pathological speech.
Brain-computer interface (BCI) is one of the tools which enables the communication between humans and devices by reflecting intention and status of humans. With the development of artificial intelligence, the interest in communication between humans and drones using electroencephalogram (EEG) is increased. Especially, in the case of controlling drone swarms such as direction or formation, there are many advantages compared with controlling a drone unit. Imagined speech is one of the endogenous BCI paradigms, which can identify intentions of users. When conducting imagined speech, the users imagine the pronunciation as if actually speaking. In contrast, overt speech is a task in which the users directly pronounce the words. When controlling drone swarms using imagined speech, complex commands can be delivered more intuitively, but decoding performance is lower than that of other endogenous BCI paradigms. We proposed the Deep-autoleaner (DAL) to learn EEG features of overt speech for imagined speech-based EEG signals classification. To the best of our knowledge, this study is the first attempt to use EEG features of overt speech to decode imagined speech-based EEG signals with an autoencoder. A total of eight subjects participated in the experiment. When classifying four words, the average accuracy of the DAL was 48.41%. In addition, when comparing the performance between w/o and w/ EEG features of overt speech, there was a performance improvement of 7.42% when including EEG features of overt speech. Hence, we demonstrated that EEG features of overt speech could improve the decoding performance of imagined speech.
Recently, there has been an increasing interest in neural speech synthesis. While the deep neural network achieves the state-of-the-art result in text-to-speech (TTS) tasks, how to generate a more emotional and more expressive speech is becoming a new challenge to researchers due to the scarcity of high-quality emotion speech dataset and the lack of advanced emotional TTS model. In this paper, we first briefly introduce and publicly release a Mandarin emotion speech dataset including 9,724 samples with audio files and its emotion human-labeled annotation. After that, we propose a simple but efficient architecture for emotional speech synthesis called EMSpeech. Unlike those models which need additional reference audio as input, our model could predict emotion labels just from the input text and generate more expressive speech conditioned on the emotion embedding. In the experiment phase, we first validate the effectiveness of our dataset by an emotion classification task. Then we train our model on the proposed dataset and conduct a series of subjective evaluations. Finally, by showing a comparable performance in the emotional speech synthesis task, we successfully demonstrate the ability of the proposed model.
A significant performance reduction is often observed in speech recognition when the rate of speech (ROS) is too low or too high. Most of present approaches to addressing the ROS variation focus on the change of speech signals in dynamic properties caused by ROS, and accordingly modify the dynamic model, e.g., the transition probabilities of the hidden Markov model (HMM). However, an abnormal ROS changes not only the dynamic but also the static property of speech signals, and thus can not be compensated for purely by modifying the dynamic model. This paper proposes an ROS learning approach based on deep neural networks (DNN), which involves an ROS feature as the input of the DNN model and so the spectrum distortion caused by ROS can be learned and compensated for. The experimental results show that this approach can deliver better performance for too slow and too fast utterances, demonstrating our conjecture that ROS impacts both the dynamic and the static property of speech. In addition, the proposed approach can be combined with the conventional HMM transition adaptation method, offering additional performance gains.
When people try to influence others to do something, they subconsciously adjust their speech to include appropriate emotional information. In order for a robot to influence people in the same way, the robot should be able to imitate the range of human emotions when speaking. To achieve this, we propose a speech synthesis method for imitating the emotional states in human speech. In contrast to previous methods, the advantage of our method is that it requires less manual effort to adjust the emotion of the synthesized speech. Our synthesizer receives an emotion vector to characterize the emotion of synthesized speech. The vector is automatically obtained from human utterances by using a speech emotion recognizer. We evaluated our method in a scenario when a robot tries to regulate an elderly person's circadian rhythm by speaking to the person using appropriate emotional states. For the target speech to imitate, we collected utterances from professional caregivers when they speak to elderly people at different times of the day. Then we conducted a subjective evaluation where the elderly participants listened to the speech samples generated by our method. The results showed that listening to the samples made the participants feel more active in the early morning and calmer in the middle of the night. This suggests that the robot may be able to adjust the participants' circadian rhythm and that the robot can potentially exert influence similarly to a person.
The enormous amount of data being generated on the web and social media has increased the demand for detecting online hate speech. Detecting hate speech will reduce their negative impact and influence on others. A lot of effort in the Natural Language Processing (NLP) domain aimed to detect hate speech in general or detect specific hate speech such as religion, race, gender, or sexual orientation. Hate communities tend to use abbreviations, intentional spelling mistakes, and coded words in their communication to evade detection, adding more challenges to hate speech detection tasks. Thus, word representation will play an increasingly pivotal role in detecting hate speech. This paper investigates the feasibility of leveraging domain-specific word embedding in Bidirectional LSTM based deep model to automatically detect/classify hate speech. Furthermore, we investigate the use of the transfer learning language model (BERT) on hate speech problem as a binary classification task. The experiments showed that domainspecific word embedding with the Bidirectional LSTM based deep model achieved a 93% f1-score while BERT achieved up to 96% f1-score on a combined balanced dataset from available hate speech datasets.
Self-supervised visual pretraining has shown significant progress recently. Among those methods, SimCLR greatly advanced the state of the art in self-supervised and semi-supervised learning on ImageNet. The input feature representations for speech and visual tasks are both continuous, so it is natural to consider applying similar objective on speech representation learning. In this paper, we propose Speech SimCLR, a new self-supervised objective for speech representation learning. During training, Speech SimCLR applies augmentation on raw speech and its spectrogram. Its objective is the combination of contrastive loss that maximizes agreement between differently augmented samples in the latent space and reconstruction loss of input representation. The proposed method achieved competitive results on speech emotion recognition and speech recognition. When used as feature extractor, our best model achieved 5.89% word error rate on LibriSpeech test-clean set using LibriSpeech 960 hours as pretraining data and LibriSpeech train-clean-100 set as fine-tuning data, which is the lowest error rate obtained in this setup to the best of our knowledge.