近日,实验室三篇论文被语音研究顶级期刊IEEE/ACM Transactions on Audio, Speech and Language Processing (TASLP)录用,一篇论文被重要期刊IEEE Signal Processing Letters (IEEE SPL)录用,论文方向涉及说话人识别中的对抗攻击、基于扩散模型的跨语种情感迁移语音合成、语音转换中基于多层级韵律建模的风格迁移、基于语言模型的语音转换。现对四篇论文工作进行简要介绍。
-1-
论文题目:Timbre-reserved Adversarial Attack in Speaker Identification 作者列表:王晴,姚继珣,张丽,郭鹏程,谢磊 发表期刊:IEEE/ACM Transactions on Audio, Speech and Language Processing 论文网址:https://arxiv.org/abs/2309.00929

论文题目:DiCLET-TTS: Diffusion Model based Cross-lingual Emotion Transfer for Text-to-Speech -- A Study between English and Mandarin 作者列表:李涛, 胡晨旭, 从坚, 朱新发, 李静北, 田乔,王玉平, 谢磊 发表期刊:IEEE/ACM Transactions on Audio, Speech and Language Processing 合作单位:字节跳动 论文网址:http://arxiv.org/abs/2309.00883

论文题目:MSM-VC: High-fidelity Source Style Transfer for Non-Parallel Voice Conversion by Multi-scale Style Modeling 作者列表:王智超,王新升,谢启聪,李涛,谢磊,田乔,王玉平 发表期刊:IEEE/ACM Transactions on Audio, Speech and Language Processing 合作单位:字节跳动 论文网址:https://arxiv.org/abs/2309.01142

论文题目:LM-VC: Zero-shot Voice Conversion via Speech Generation based on Language Models 作者列表:王智超,陈远哲,谢磊,田乔,王玉平 发表期刊:IEEE Signal Processing Letters 合作单位:字节跳动 论文网址:https://arxiv.org/abs/2306.10521

