Jump to Content
Matt Shannon

Matt Shannon

Researcher in speech synthesis, speech recognition, generative modeling, machine learning.
Authored Publications
Google Publications
Other Publications
Sort By
  • Title
  • Title, descending
  • Year
  • Year, descending
    Preview abstract This work explores the task of synthesizing speech in human-sounding voices unseen in any training set. We call this task "speaker generation", and present TacoSpawn, a system that performs competitively at this task. TacoSpawn is a deep generative text-to-speech model that learns a distribution over a speaker embedding space, which enables sampling of novel and diverse speakers. Our method is easy to implement, and does not require transfer learning from speaker ID systems. We present objective and subjective metrics for evaluating performance on this task, and demonstrate that our proposed objective metrics correlate with human perception of speaker similarity. View details
    Preview abstract Recent work has explored sequence-to-sequence latent variable models for expressive speech synthesis (supporting control and transfer of prosody and style), but has not presented a coherent framework for understanding the trade-offs between the competing methods. In this paper, we propose embedding capacity (the amount of information the embedding contains about the data) as a unified method of analyzing the behavior of latent variable models of speech, comparing existing heuristic (non-variational) methods to variational methods that are able to explicitly constrain capacity using an upper bound on representational mutual information. In our proposed model (Capacitron), we show that by adding conditional dependencies to the variational posterior such that it matches the form of the true posterior, the same model can be used for high-precision prosody transfer, text-agnostic style transfer, and generation of natural-sounding prior samples. For multi-speaker models, Capacitron is able to preserve target speaker identity during inter-speaker prosody transfer and when drawing samples from the latent prior. Lastly, we introduce a method for decomposing embedding capacity hierarchically across two sets of latents, allowing a portion of the latent variability to be specified and the remaining variability sampled from a learned prior. Audio examples are available on the web. View details
    Preview abstract We present a novel generative model that combines state-of-the-art neural text-to-speech (TTS) with semi-supervised probabilistic latent variable models. By providing partial supervision to some of the latent variables, we are able to force them to take on consistent and interpretable purposes, which previously hasn't been possible with purely unsupervised methods. We demonstrate that our model is able to reliably discover and control important but rarely labelled attributes of speech, such as affect and speaking rate, with as little as 0.5\% (15 minutes) supervision. Even at such low supervision levels we do not observe a degradation of synthesis quality compared to a state-of-the-art baseline. View details
    Preview abstract Despite the ability to produce human-level speech for in-domain text, attention-based end-to-end text-to-speech (TTS) systems suffer from text alignment failures that increase in frequency for out-of-domain text. We show that these failures can be addressed using simple location-relative attention mechanisms that do away with content-based query/key comparisons. We compare two families of attention mechanisms: location-relative GMM-based mechanisms and additive energy-based mechanisms. We suggest simple modifications to GMM-based attention that allow it to align quickly and consistently during training, and introduce a new location-relative attention mechanism to the additive energy-based family, called Dynamic Convolution Attention (DCA). We compare the various mechanisms in terms of alignment speed and consistency during training, naturalness, and ability to generalize to long utterances, and conclude that GMM attention and DCA can generalize to very long utterances, while preserving naturalness for shorter, in-domain utterances. View details
    Preview abstract State-level minimum Bayes risk (sMBR) training has become the de facto standard for sequence-level training of speech recognition acoustic models. It has an elegant formulation using the expectation semiring, and gives large improvements in word error rate (WER) over models trained solely using cross-entropy (CE) or connectionist temporal classification (CTC). sMBR training optimizes the expected number of frames at which the reference and hypothesized acoustic states differ. It may be preferable to optimize the expected WER, but WER does not interact well with the expectation semiring, and previous approaches based on computing expected WER exactly involve expanding the lattices used during training. In this paper we show how to perform optimization of the expected WER by sampling paths from the lattices used during conventional sMBR training. The gradient of the expected WER is itself an expectation, and so may be approximated using Monte Carlo sampling. We show experimentally that optimizing WER during acoustic model training gives 5% relative improvement in WER over a well-tuned sMBR baseline on a 2-channel query recognition task (Google Home). View details
    Preview abstract This paper describes the technical and system building advances made to the Google Home multichannel speech recognition system, which was launched in November 2016. Technical advances include an adaptive dereverberation frontend, the use of neural network models that do multichannel processing jointly with acoustic modeling, and grid lstms to model frequency variations. On the system level, improvements include adapting the model using Google Home specific data. We present results on a variety of multichannel sets. The combination of technical and system advances result in a reduction of WER of over 18\% relative compared to the current production system. View details
    Preview abstract In many streaming speech recognition applications such as voice search it is important to determine quickly and accurately when the user has finished speaking their query. A conventional approach to this task is to declare end-of-query whenever a fixed interval of silence is detected by a voice activity detector (VAD) trained to classify each frame as speech or silence. However silence detection and end-of-query detection are fundamentally different tasks, and the criterion used during VAD training may not be optimal. In particular the conventional approach ignores potential acoustic cues such as filler sounds and past speaking rate which may indicate whether a given pause is temporary or query-final. In this paper we present a simple modification to make the conventional VAD training criterion more closely related to end-of-query detection. A unidirectional long short-term memory architecture allows the system to remember past acoustic events, and the training criterion incentivizes the system to learn to use any acoustic cues relevant to predicting future user intent. We show experimentally that this approach improves latency at a given accuracy for end-of-query detection for voice search. View details
    No Results Found