DQR-TTS: Semi-supervised Text-to-speech Synthesis with Dynamic Quantized Representation | IEEE Conference Publication | IEEE Xplore