2018 · 285 citations · 33 references
ComputingComputer ScienceArtificial IntelligenceMachine LearningNatural Language ProcessingDeep LearningSpeech RecognitionSpeech SynthesisText GenerationConvolutional Neural NetworksNeural Speech SynthesisRecurrent Neural NetworksText-to-speech Synthesis
Recurrent neural networks dominate sequential modeling in TTS, yet recent CNN-based synthesis shows faster, more parallelizable alternatives. This work aims to prove that a CNN-only TTS system can be trained efficiently, reducing the high computational costs of RNN-based approaches. The authors employ a deep convolutional neural network architecture without recurrent units, guided by attention mechanisms, to synthesize speech. In experiments, the CNN TTS was trained overnight (15 h) on a standard gaming PC with two GPUs, producing speech of almost acceptable quality.
This paper describes a novel text-to-speech (TTS) technique based on deep convolutional neural networks (CNN), without use of any recurrent units. Recurrent neural networks (RNN) have become a standard technique to model sequential data recently, and this technique has been used in some cutting-edge neural TTS techniques. However, training RNN components often requires a very powerful computer, or a very long time, typically several days or weeks. Recent other studies, on the other hand, have shown that CNN-based sequence synthesis can be much faster than RNN-based techniques, because of high parallelizability. The objective of this paper is to show that an alternative neural TTS based only on CNN alleviate these economic costs of training. In our experiment, the proposed Deep Convolutional TTS was sufficiently trained overnight (15 hours), using an ordinary gaming PC equipped with two GPUs, while the quality of the synthesized speech was almost acceptable.
33
Adam: A Method for Stochastic Optimization
Diederik P. Kingma, Jimmy Ba · UvA-DARE (University of Amsterdam) · 2014 · 84.8K citations · Full text
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
Kaiming He, Xiangyu Zhang, Shaoqing Ren et al. · 2015 · 18.8K citations
Convolutional Neural Networks for Sentence Classification