Towards Lifelong Learning of Multilingual Text-To-Speech Synthesis
This addresses the storage and computation burden for multilingual TTS systems, but it is incremental as it builds on existing lifelong learning techniques.
This work tackled the problem of catastrophic forgetting in multilingual Text-To-Speech synthesis by proposing a lifelong learning approach with a data-replay-based method, achieving a 43% reduction in Mel-Cepstral Distortion compared to a fine-tuning baseline.
This work presents a lifelong learning approach to train a multilingual Text-To-Speech (TTS) system, where each language was seen as an individual task and was learned sequentially and continually. It does not require pooled data from all languages altogether, and thus alleviates the storage and computation burden. One of the challenges of lifelong learning methods is "catastrophic forgetting": in TTS scenario it means that model performance quickly degrades on previous languages when adapted to a new language. We approach this problem via a data-replay-based lifelong learning method. We formulate the replay process as a supervised learning problem, and propose a simple yet effective dual-sampler framework to tackle the heavily language-imbalanced training samples. Through objective and subjective evaluations, we show that this supervised learning formulation outperforms other gradient-based and regularization-based lifelong learning methods, achieving 43% Mel-Cepstral Distortion reduction compared to a fine-tuning baseline.