Yujie Tu

Yujie Tu

个人简介About me

大家好，我是涂宇杰。目前为中国科学院自动化研究所和上海创智学院的联培博士生，由陈谐副教授与张家俊研究员共同指导。我出生于江苏无锡，现居上海。我本科毕业于中国科学院大学。此外，我曾以访问学生身份在美国加州大学伯克利分校学习半年。

我很幸运能在求学过程中走进许多顶尖科研机构。自 2024 年 3 月起，我成为上海交通大学跨媒体语言智能实验室（X-LANCE）的科研实习生，由陈谐副教授指导。目前，我在微软亚洲研究院（MSRA）的通用人工智能研究组实习，由余剑威博士与彭智亮博士共同指导。在本科的科研实践与毕业设计阶段，我的指导老师是刘斌副研究员。在此之前，我也曾在字节跳动的 TikTok 部门进行过暑期实习。

我目前的主要研究方向集中在语音识别，具体包括多说话人语音识别、低资源语音识别、流式语音识别等。此外，我对其他语音方向也有潜在的兴趣。

Hi, I am Yujie Tu, a Ph.D. student jointly trained by the Institute of Automation, Chinese Academy of Sciences and the Shanghai Innovation Institute, advised by Assoc. Prof. Xie Chen and Prof. Jiajun Zhang. I was born in Wuxi, Jiangsu, and now live in Shanghai. I received my bachelor’s degree from the University of Chinese Academy of Sciences, and spent half a year at the University of California, Berkeley as a visiting student.

I have been fortunate to work in several outstanding research groups. Since March 2024 I have been a research intern at the Cross Media Language Intelligence Lab (X-LANCE), Shanghai Jiao Tong University, advised by Assoc. Prof. Xie Chen. I am currently interning with the General Artificial Intelligence group at Microsoft Research Asia (MSRA), advised by Dr. Jianwei Yu and Dr. Zhiliang Peng. My undergraduate research and thesis were supervised by Assoc. Prof. Bin Liu. Before that I spent a summer interning on TikTok at ByteDance.

My research centres on speech recognition — multi-speaker ASR, low-resource ASR, and streaming ASR in particular. I am also drawn to the wider speech field.

近期动态News

  1. 最新New VibeVoice-ASR-BitNet 技术报告发布，在边缘 CPU 上实现实时多语种识别。VibeVoice-ASR-BitNet technical report is out — real-time multilingual recognition on edge CPUs.
  2. 最新New 第一作者工作 GigaSpeechBench 发布，一个 680 小时的真实场景多语种语音评测基准。GigaSpeechBench is out — my first-author work, a 680-hour real-world multilingual speech benchmark.
  3. 最新New OpenSTBench 发布，面向语音翻译的多维度统一评测框架。OpenSTBench is out — a unified multidimensional evaluation framework for speech translation.
  4. VibeVoice-ASR 技术报告发布，单次处理长达 60 分钟音频的通用语音理解框架。VibeVoice-ASR technical report is out — single-pass speech understanding for up to 60 minutes of audio.
更早的 3 条动态3 earlier updates
  1. Habibi 发布，首个开源的统一多方言阿拉伯语语音合成框架。Habibi is out — the first open-source unified-dialectal Arabic speech synthesis framework.
  2. 加入微软亚洲研究院（MSRA）通用人工智能研究组实习。Started an internship with the General Artificial Intelligence group at Microsoft Research Asia.
  3. 在字节跳动 TikTok 部门进行暑期实习。Summer internship on TikTok at ByteDance.

研究方向Research areas

2 篇works

低资源语音技术Low-Resource Speech Technologies

为常规评测长期忽视的语言、方言与人群，构建基准数据与识别、合成等语音技术。Benchmarks, recognition and synthesis for languages, dialects, and communities that conventional evaluation leaves out.

 2 篇works

多说话人语音识别Multi-Speaker Speech Recognition

把转写、说话人归属与实际部署统一起来的长音频高效语音识别系统。Long-form and efficient ASR systems that unify transcription, speaker attribution, and practical deployment.

 1 篇works

语音合成、翻译和其他语音技术Speech Synthesis, Translation and Other Speech Technologies

语音翻译与合成系统，以及它们所需的评测方法，兼及相邻的语音技术方向。Speech translation and synthesis systems, and the evaluation protocols they need, alongside neighbouring speech technologies.

精选论文Selected publications

查看全部论文 →All publications →

教育经历Education

  • 2026.9 – 至今 · 上海创智学院 & 中国科学院自动化研究所，模式识别与智能系统
  • 2025.1 – 2025.5 · 加州大学伯克利分校，访问学生
  • 2022.9 – 2026.6 · 中国科学院大学，人工智能学院
  • 2019.9 – 2022.6 · 江苏省锡山高级中学
  • Sep 2026 – present · Shanghai Innovation Institute & Institute of Automation, CAS — Pattern Recognition and Intelligent Systems
  • Jan 2025 – May 2025 · University of California, Berkeley — visiting student
  • Sep 2022 – Jun 2026 · University of Chinese Academy of Sciences — School of Artificial Intelligence
  • Sep 2019 – Jun 2022 · Xishan Senior High School of Jiangsu Province

工作经历Experience

  • 2025.10 – 至今 · 微软亚洲研究院，通用人工智能研究组
  • 2025.6 – 2025.9 · 字节跳动，TikTok
  • Oct 2025 – present · Microsoft Research Asia — General Artificial Intelligence group
  • Jun 2025 – Sep 2025 · ByteDance — TikTok

奖项荣誉Awards

  • 中国科学院大学 二等学业奖学金
  • 中国科学院大学 三等学业奖学金
  • Second-Class Academic Scholarship, University of Chinese Academy of Sciences
  • Third-Class Academic Scholarship, University of Chinese Academy of Sciences