In this paper, we propose the use of self-supervised pretraining on a large unlabelled data set to improve the performance of a personalized voice activity detection (VAD) model in adverse conditions. We pretrain a long short-term memory (LSTM)-encoder using the autoregressive predictive coding (APC) framework and fine-tune it for personalized VAD. We also propose a denoising variant of APC, with the goal of improving the robustness of personalized VAD. The trained models are systematically evaluated on both clean speech and speech contaminated by various types of noise at different SNR-levels and compared to a purely supervised model. Our experiments show that self-supervised pretraining not only improves performance in clean conditions, but also yields models which are more robust to adverse conditions compared to purely supervised learning.
* To be published at ICASSP2024, 14th of April 2024, Seoul, South
Korea. Copyright (c) 2023 IEEE. 5 pages, 2, figures, 5 tables
In this paper, we derive a PAC-Bayes bound on the generalisation gap, in a supervised time-series setting for a special class of discrete-time non-linear dynamical systems. This class includes stable recurrent neural networks (RNN), and the motivation for this work was its application to RNNs. In order to achieve the results, we impose some stability constraints, on the allowed models. Here, stability is understood in the sense of dynamical systems. For RNNs, these stability conditions can be expressed in terms of conditions on the weights. We assume the processes involved are essentially bounded and the loss functions are Lipschitz. The proposed bound on the generalisation gap depends on the mixing coefficient of the data distribution, and the essential supremum of the data. Furthermore, the bound converges to zero as the dataset size increases. In this paper, we 1) formalize the learning problem, 2) derive a PAC-Bayesian error bound for such systems, 3) discuss various consequences of this error bound, and 4) show an illustrative example, with discussions on computing the proposed bound. Unlike other available bounds the derived bound holds for non i.i.d. data (time-series) and it does not grow with the number of steps of the RNN.
Diffusion models are a new class of generative models that have shown outstanding performance in image generation literature. As a consequence, studies have attempted to apply diffusion models to other tasks, such as speech enhancement. A popular approach in adapting diffusion models to speech enhancement consists in modelling a progressive transformation between the clean and noisy speech signals. However, one popular diffusion model framework previously laid in image generation literature did not account for such a transformation towards the system input, which prevents from relating the existing diffusion-based speech enhancement systems with the aforementioned diffusion model framework. To address this, we extend this framework to account for the progressive transformation between the clean and noisy speech signals. This allows us to apply recent developments from image generation literature, and to systematically investigate design aspects of diffusion models that remain largely unexplored for speech enhancement, such as the neural network preconditioning, the training loss weighting, the stochastic differential equation (SDE), or the amount of stochasticity injected in the reverse process. We show that the performance of previous diffusion-based speech enhancement systems cannot be attributed to the progressive transformation between the clean and noisy speech signals. Moreover, we show that a proper choice of preconditioning, training loss weighting, SDE and sampler allows to outperform a popular diffusion-based speech enhancement system in terms of perceptual metrics while using fewer sampling steps, thus reducing the computational cost by a factor of four.
Diffusion models are a new class of generative models that have recently been applied to speech enhancement successfully. Previous works have demonstrated their superior performance in mismatched conditions compared to state-of-the art discriminative models. However, this was investigated with a single database for training and another one for testing, which makes the results highly dependent on the particular databases. Moreover, recent developments from the image generation literature remain largely unexplored for speech enhancement. These include several design aspects of diffusion models, such as the noise schedule or the reverse sampler. In this work, we systematically assess the generalization performance of a diffusion-based speech enhancement model by using multiple speech, noise and binaural room impulse response (BRIR) databases to simulate mismatched acoustic conditions. We also experiment with a noise schedule and a sampler that have not been applied to speech enhancement before. We show that the proposed system substantially benefits from using multiple databases for training, and achieves superior performance compared to state-of-the-art discriminative models in both matched and mismatched conditions. We also show that a Heun-based sampler achieves superior performance at a smaller computational cost compared to a sampler commonly used for speech enhancement.
We consider speech enhancement for signals picked up in one noisy environment that must be rendered to a listener in another noisy environment. For both far-end noise reduction and near-end listening enhancement, it has been shown that excessive focus on noise suppression or intelligibility maximization may lead to excessive speech distortions and quality degradations in favorable noise conditions, where intelligibility is already at ceiling level. Recently [1,2] propose to remedy this with a minimum processing framework that either reduces noise or enhances listening a minimum amount given that a certain intelligibility criterion is still satisfied. Additionally, it has been shown that joint consideration of both environments improves speech enhancement performance. In this paper, we formulate a joint far- and near-end minimum processing framework, that improves intelligibility while limiting speech distortions in favorable noise conditions. We provide closed-form solutions to specific boundary scenarios and investigate performance for the general case using numerical optimization. We also show that concatenating existing minimum processing far- and near-end enhancement methods preserves the effects of the initial methods. Results show that the joint optimization can further improve performance compared to the concatenated approach.
Several recent works have adapted Masked Autoencoders (MAEs) for learning general-purpose audio representations. However, they do not address two key aspects of modelling multi-domain audio data: (i) real-world audio tasks consist of a combination of local+global contexts, and (ii) real-world audio signals are complex compositions of several acoustic elements with different time-frequency characteristics. To address these concerns, this work proposes a Multi-Window Masked Autoencoder (MW-MAE) fitted with a novel Multi-Window Multi-Head Attention module that can capture information at multiple local and global contexts in every decoder transformer block through attention heads of several distinct local and global windows. Empirical results on ten downstream audio tasks show that MW-MAEs consistently outperform standard MAEs in overall performance and learn better general-purpose audio representations, as well as demonstrate considerably better scaling characteristics. Exploratory analyses of the learned representations reveals that MW-MAE encoders learn attention heads with more distinct entropies compared to those learned by MAEs, while attention heads across the different transformer blocks in MW-MAE decoders learn correlated feature representations, enabling each block to independently capture local and global information, leading to a decoupled feature hierarchy. Code for feature extraction and downstream experiments along with pre-trained weights can be found at https://github.com/10997NeurIPS23/10997_mwmae.
Audio and visual modalities are inherently connected in speech signals: lip movements and facial expressions are correlated with speech sounds. This motivates studies that incorporate the visual modality to enhance an acoustic speech signal or even restore missing audio information. Specifically, this paper focuses on the problem of audio-visual speech inpainting, which is the task of synthesizing the speech in a corrupted audio segment in a way that it is consistent with the corresponding visual content and the uncorrupted audio context. We present an audio-visual transformer-based deep learning model that leverages visual cues that provide information about the content of the corrupted audio. It outperforms the previous state-of-the-art audio-visual model and audio-only baselines. We also show how visual features extracted with AV-HuBERT, a large audio-visual transformer for speech recognition, are suitable for synthesizing speech.
In this paper we derive a Probably Approxilmately Correct(PAC)-Bayesian error bound for linear time-invariant (LTI) stochastic dynamical systems with inputs. Such bounds are widespread in machine learning, and they are useful for characterizing the predictive power of models learned from finitely many data points. In particular, with the bound derived in this paper relates future average prediction errors with the prediction error generated by the model on the data used for learning. In turn, this allows us to provide finite-sample error bounds for a wide class of learning/system identification algorithms. Furthermore, as LTI systems are a sub-class of recurrent neural networks (RNNs), these error bounds could be a first step towards PAC-Bayesian bounds for RNNs.
* arXiv admin note: text overlap with arXiv:2212.14838
In this paper we derive a PAC-Bayesian-Like error bound for a class of stochastic dynamical systems with inputs, namely, for linear time-invariant stochastic state-space models (stochastic LTI systems for short). This class of systems is widely used in control engineering and econometrics, in particular, they represent a special case of recurrent neural networks. In this paper we 1) formalize the learning problem for stochastic LTI systems with inputs, 2) derive a PAC-Bayesian-Like error bound for such systems, 3) discuss various consequences of this error bound.
In the context of keyword spotting (KWS), the replacement of handcrafted speech features by learnable features has not yielded superior KWS performance. In this study, we demonstrate that filterbank learning outperforms handcrafted speech features for KWS whenever the number of filterbank channels is severely decreased. Reducing the number of channels might yield certain KWS performance drop, but also a substantial energy consumption reduction, which is key when deploying common always-on KWS on low-resource devices. Experimental results on a noisy version of the Google Speech Commands Dataset show that filterbank learning adapts to noise characteristics to provide a higher degree of robustness to noise, especially when dropout is integrated. Thus, switching from typically used 40-channel log-Mel features to 8-channel learned features leads to a relative KWS accuracy loss of only 3.5% while simultaneously achieving a 6.3x energy consumption reduction.