IEEE TASLP Article

Human speech can be characterized by different components, including semantic content, speaker identity and prosodic information. Significant progress has been made in disentangling representations for semantic content and speaker identity in speech recognition and speaker verification tasks respectively. However, it is still an open challenging question to extract prosodic information because of the intrinsic association of different attributes, such as timbre and rhythm, and because of the need for supervised training schemes to achieve robust speech recognition.

Speech applications in far-field real world settings often deal with signals that are corrupted by reverberation. The task of dereverberation constitutes an important step to improve the audible quality and to reduce the error rates in applications like automatic speech recognition (ASR). We propose a unified framework of speech dereverberation for improving the speech quality and the ASR performance using the approach of envelope-carrier decomposition provided by an autoregressive (AR) model.

Question answering requiring numerical reasoning, which generally involves symbolic operations such as sorting, counting, and addition, is a challenging task. To address such a problem, existing mixture-of-experts (MoE)-based methods design several specific answer predictors to handle different types of questions and achieve promising performance. However, they ignore the modeling and exploitation of fine-grained reasoning-related operations to support numerical reasoning, encountering the inadequacy in reasoning capability and interpretability.

The speaker recognition evaluation is conducted in a framework in which three score distributions and two decision thresholds are employed, and the statistic of interest is an average of the two weighted sums of the probabilities of type I and type II errors at the two thresholds correspondingly. And data dependence caused by multiple use of the same subjects exists ubiquitously in order to generate more samples because of limited resources.

This study proposes a cross-domain multi-objective speech assessment model, called MOSA-Net, which can simultaneously estimate the speech quality, intelligibility, and distortion assessment scores of an input speech signal. MOSA-Net comprises a convolutional neural network and bidirectional long short-term memory architecture for representation extraction, and a multiplicative attention layer and a fully connected layer for each assessment metric prediction. Additionally, cross-domain features (spectral and time-domain features) and latent representations from self-supervised learned (SSL) models are used as inputs to combine rich acoustic information to obtain more accurate assessments.

This paper introduces a new framework for non-parallel emotion conversion in speech. Our framework is based on two key contributions. First, we propose a stochastic version of the popular Cycle-GAN model. Our modified loss function introduces a Kullback–Leibler (KL) divergence term that aligns the source and target data distributions learned by the generators, thus overcoming the limitations of sample-wise generation. By using a variational approximation to this stochastic loss function, we show that our KL divergence term can be implemented via a paired density discriminator.

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;

Emotional voice conversion (VC) aims to convert a neutral voice to an emotional one while retaining the linguistic information and speaker identity. We note that the decoupling of emotional features from other speech information (such as content, speaker identity, etc.) is the key to achieving promising performance. Some recent attempts of speech representation decoupling on the neutral speech cannot work well on the emotional speech, due to the more complex entanglement of acoustic properties in the latter. 

Detection of speech and music signals in isolated and overlapped conditions is an essential preprocessing step for many audio applications. Speech signals have wavy and continuous harmonics, while music signals exhibit horizontally linear and discontinuous harmonic patterns. Music signals also contain more percussive components than speech signals, manifested as vertical striations in the spectrograms.

Deep neural networks (DNNs) represent the mainstream methodology for supervised speech enhancement, primarily due to their capability to model complex functions using hierarchical representations. However, a recent study revealed that DNNs trained on a single corpus fail to generalize to untrained corpora, especially in low signal-to-noise ratio (SNR) conditions.