Ph.D. Candidate
NTU EECS

Xuanjun Chen (陳炫均)

Ph.D. Candidate, NTU EECS · Taipei, Taiwan

I am a Ph.D. student at National Taiwan University (NTU), advised by Hung-yi Lee and Jyh-Shing Roger Jang. I received my M.S. from NTU (2023) and B.S. from Taiwan Tech (2020). Recently, my research focuses on speech tokenization and spoken language modeling.

My previous research encompassed speech processing, speech/text LLMs, and deepfake detection. I lead Deepfake research by proposing SingGraph for SingFake detection, developing CodecFake+ for data scarcity, and pioneering source tracing with SASTNet. Beyond security, I am a core contributor to several audio LLM and benchmarking projects, including DeSTA 2.5, Dynamic-SUPERB Phase 2, and Codec-SUPERB. I also mentor research on optimizing RAG efficiency for LLMs and uncovering their practical limitations. These works earned several honors, including best paper awards & travel grants. (see Honors).

[ Email | Linkedin | Twitter | Github | Google Scholar ]

Selected Publications (* equal contribution)

[1]

CodecFake+: Codec-Based Resynthesized Data as a Proxy for Detecting CodecFake Speech
Xuanjun Chen*, Jiawei Du*, Haibin Wu, Lin Zhang, I-Ming Lin, ..., Jyh-Shing Roger Jang, Hung-yi Lee

TASLP 2026 [ bib | arXiv | Project Page | Hugging Face | Code ]

@misc{chen2025codecfake,
    title={CodecFake+: A Large-Scale Neural Audio Codec-Based Deepfake Speech Dataset},
    author={Xuanjun Chen and Jiawei Du and Haibin Wu and Lin Zhang and I-Ming Lin and I-Hsiang Chiu and Wenze Ren and Yuan Tseng and Yu Tsao and Jyh-Shing Roger Jang and Hung-yi Lee},
    year={2025},
    eprint={2501.08238},
    archivePrefix={arXiv},
    primaryClass={cs.SD}
}

[2]

DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment
Ke-Han Lu, Zhehuai Chen, Szu-Wei Fu, ..., Xuanjun Chen, ..., Boris Ginsburg, Yu-Chiang Frank Wang, Hung-yi Lee

TASLP 2026 [ bib | arXiv | Code ]

@misc{lu2025desta25,
    title={DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment},
    author={Ke-Han Lu and Zhehuai Chen and Szu-Wei Fu and Chao-Han Huck Yang and Sung-Feng Huang and Chih-Kai Yang and Chee-En Yu and Chun-Wei Chen and Wei-Chih Chen and Chien-yu Huang and Yi-Cheng Lin and Yu-Xiang Lin and Chi-An Fu and Chun-Yi Kuan and Wenze Ren and Xuanjun Chen and Wei-Ping Huang and En-Pei Hu and Tzu-Quan Lin and Yuan-Kuei Wu and Kuan-Po Huang and Hsiao-Ying Huang and Huang-Cheng Chou and Kai-Wei Chang and Cheng-Han Chiang and Boris Ginsburg and Yu-Chiang Frank Wang and Hung-yi Lee},
    year={2025},
    eprint={2507.02768},
    archivePrefix={arXiv},
    primaryClass={cs.SD}
}

[3]

Joint Fullband-Subband Modeling for High-Resolution SingFake Detection
Xuanjun Chen*, Chia-Yu Hu*, Sung-Feng Huang, Haibin Wu, Hung-yi Lee, Jyh-Shing Roger Jang

INTERSPEECH 2026 (Submitted) [ bib | arXiv ]

@article{chen2026joint,
  title={Joint Fullband-Subband Modeling for High-Resolution SingFake Detection},
  author={Chen, Xuanjun and Hu, Chia-Yu and Huang, Sung-Feng and Wu, Haibin and Lee, Hung-yi and Jang, Jyh-Shing Roger},
  journal={arXiv preprint arXiv:2604.04841},
  year={2026}
}

[4]

How Does Instrumental Music Help SingFake Detection?
Xuanjun Chen, Chia-Yu Hu, I-Ming Lin, Yi-Cheng Lin, I-Hsiang Chiu, ..., Hung-yi Lee, Jyh-Shing Roger Jang

ICASSP 2026 [ bib | arXiv ]

@misc{chen2025how,
    title={How Does Instrumental Music Help SingFake Detection?},
    author={Xuanjun Chen and Chia-Yu Hu and I-Ming Lin and Yi-Cheng Lin and I-Hsiang Chiu and You Zhang and Sung-Feng Huang and Yi-Hsuan Yang and Haibin Wu and Hung-yi Lee and Jyh-Shing Roger Jang},
    year={2025},
    eprint={2509.14675},
    archivePrefix={arXiv},
    primaryClass={cs.SD}
}

[5]

Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the Capabilities of Spoken Language Models with 180 Tasks
Chien-yu Huang, Wei-Chih Chen, Shu-wen Yang, ..., Xuanjun Chen, ..., Shinji Watanabe, Hung-yi Lee

ICLR 2025 [ bib | arXiv | OpenReview | Code ]

@inproceedings{huang2025dynamic,
    title={Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the <br> Capabilities of Spoken Language Models with 180 Tasks},
    author={Chien-yu Huang and Wei-Chih Chen and Shu-wen Yang and Andy T. Liu and Chen-An Li and Yu-Xiang Lin and Wei-Cheng Tseng and Anuj Diwan and Yi-Jen Shih and Jiatong Shi and William Chen and Chih-Kai Yang and Wenze Ren and Xuanjun Chen and Chi-Yuan Hsiao and Puyuan Peng and Shih-Heng Wang and Chun-Yi Kuan and Ke-Han Lu and Kai-Wei Chang and Fabian Ritter-Gutierrez and Kuan-Po Huang and Siddhant Arora and You-Kuan Lin and Ming To Chuang and Eunjung Yeo and Kalvin Chang and Chung-Ming Chien and Kwanghee Choi and Jun-You Wang and Cheng-Hsiu Hsieh and Yi-Cheng Lin and Chee-En Yu and I-Hsiang Chiu and Heitor R. Guimarães and Jionghao Han and Tzu-Quan Lin and Tzu-Yuan Lin and Homu Chang and Ting-Wu Chang and Chun Wei Chen and Shou-Jen Chen and Yu-Hua Chen and Hsi-Chun Cheng and Kunal Dhawan and Jia-Lin Fang and Shi-Xin Fang and Kuan-Yu Fang Chiang and Chi An Fu and Hsien-Fu Hsiao and Ching Yu Hsu and Shao-Syuan Huang and Lee Chen Wei and Hsi-Che Lin and Hsuan-Hao Lin and Hsuan-Ting Lin and Jian-Ren Lin and Ting-Chun Liu and Li-Chun Lu and Tsung-Min Pai and Ankita Pasad and Shih-Yun Shan Kuan and Suwon Shon and Yuxun Tang and Yun-Shao Tsai and Jui-Chiang Wei and Tzu-Chieh Wei and Chengxi Wu and Dien-Ruei Wu and Chao-Han Huck Yang and Chieh-Chi Yang and Jia Qi Yip and Shao-Xiang Yuan and Vahid Noroozi and Zhehuai Chen and Haibin Wu and Karen Livescu and David Harwath and Shinji Watanabe and Hung-yi Lee},
    booktitle={International Conference on Learning Representations},
    year={2025},
    url={https://openreview.net/forum?id=s7lzZpAW7T}
}

[6]

A Preliminary Study of RAG for Taiwanese Historical Archives
Claire Lin*, Bo-Han Feng*, Xuanjun Chen*, Te-Lun Yang, Hung-yi Lee, Jyh-Shing Roger Jang

ROCLING 2025 [ bib | arXiv | ACL Anthology ] Best Paper Award

@inproceedings{lin2025preliminary,
    title = "A Preliminary Study of {RAG} for {T}aiwanese Historical Archives",
    author = "Lin, Claire  and Feng, Bo-Han  and Chen, Xuanjun  and Yang, Te-Lun  and Lee, Hung-yi  and Jang, Jyh-Shing Roger",
    booktitle = "Proceedings of the 37th Conference on Computational Linguistics and Speech Processing (ROCLING 2025)",
    year = "2025",
    url = "https://aclanthology.org/2025.rocling-main.6"
}

[7]

Towards Generalized Source Tracing for Codec-Based Deepfake Speech
Xuanjun Chen*, I-Ming Lin*, Lin Zhang, Haibin Wu, Hung-yi Lee, Jyh-Shing Roger Jang

IEEE ASRU 2025 [ bib | arXiv | Code ] Best Student Paper nominee

@inproceedings{chen2025towards,
    title={Towards Generalized Source Tracing for Codec-Based Deepfake Speech},
    author={Xuanjun Chen and I-Ming Lin and Lin Zhang and Haibin Wu and Hung-yi Lee and Jyh-Shing Roger Jang},
    booktitle={2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)},
    year={2025}
}

[8]

Codec-Based Deepfake Source Tracing via Neural Audio Codec Taxonomy
Xuanjun Chen*, I-Ming Lin*, Lin Zhang, Jiawei Du, Haibin Wu, Hung-yi Lee, Jyh-Shing Roger Jang

INTERSPEECH 2025 [ bib | arXiv | ISCA | Code ]

@inproceedings{chen2025codecbased,
    title={Codec-Based Deepfake Source Tracing via Neural Audio Codec Taxonomy},
    author={Xuanjun Chen and I-Ming Lin and Lin Zhang and Jiawei Du and Haibin Wu and Hung-yi Lee and Jyh-Shing Roger Jang},
    booktitle={Interspeech 2025},
    year={2025},
    url={https://www.isca-archive.org/interspeech_2025/chen25j_interspeech.pdf}
}

[9]

Singing Voice Graph Modeling for SingFake Detection
Xuanjun Chen, Haibin Wu, Jyh-Shing Roger Jang, Hung-yi Lee

INTERSPEECH 2024 (Oral) [ bib | arXiv | ISCA | Code | Lightning Talk ]

@inproceedings{chen2024singing,
    title={Singing Voice Graph Modeling for SingFake Detection},
    author={Xuanjun Chen and Haibin Wu and Jyh-Shing Roger Jang and Hung-yi Lee},
    booktitle={Interspeech 2024},
    year={2024},
    url={https://www.isca-archive.org/interspeech_2024/chen24o_interspeech.pdf}
}

[10]

Codec-SUPERB @ SLT 2024: A lightweight benchmark for neural codec models
Haibin Wu, Xuanjun Chen, Yi-Cheng Lin, Jiawei Du, Kai-Wei Chang, ..., Shinji Watanabe, Hung-yi Lee

IEEE SLT 2024 [ bib | arXiv | IEEE Xplore ]

@inproceedings{wu2024codecsuperbslt,
    title={Codec-SUPERB @ SLT 2024: A lightweight benchmark for neural codec models},
    author={Haibin Wu and Xuanjun Chen and Yi-Cheng Lin and Jiawei Du and Kai-Wei Chang and Ke-Han Lu and Alexander Liu and Ho-Lam Chung and Yuan-Kuei Wu and Dongchao Yang and Songxiang Liu and Yi-Chiao Wu and Xu Tan and James Glass and Shinji Watanabe and Hung-yi Lee},
    booktitle={2024 IEEE Spoken Language Technology Workshop (SLT)},
    year={2024}
}

[11]

Codec-SUPERB: An In-Depth Analysis of Sound Codec Models
Haibin Wu, Ho-Lam Chung, Yi-Cheng Lin, ..., Xuanjun Chen, ..., Kai-Wei Chang, Alexander H. Liu, Hung-yi Lee

@inproceedings{wu-etal-2024-codec,
    title = "Codec-{SUPERB}: An In-Depth Analysis of Sound Codec Models",
    author = "Wu, Haibin  and Chung, Ho-Lam  and Lin, Yi-Cheng  and Wu, Yuan-Kuei  and Chen, Xuanjun  and Pai, Yu-Chi  and Wang, Hsiu-Hsuan  and Chang, Kai-Wei  and Liu, Alexander  and Lee, Hung-yi",
    editor = "Ku, Lun-Wei  and Martins, Andre  and Srikumar, Vivek",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2024",
    year = "2024",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.findings-acl.616/",
    doi = "10.18653/v1/2024.findings-acl.616",
    pages = "10330--10348",
}

[12]

Towards audio language modeling-an overview
Haibin Wu, Xuanjun Chen, Yi-Cheng Lin, Kai-wei Chang, Ho-Lam Chung, Alexander Liu, Hung-yi Lee

Tech Report, Feb. 2024 [ bib | arXiv | Awesome ]

@misc{wu2024towards,
    title={Towards audio language modeling - an overview},
    author={Haibin Wu and Xuanjun Chen and Yi-Cheng Lin and Kai-wei Chang and Ho-Lam Chung and Alexander Liu and Hung-yi Lee},
    year={2024},
    eprint={2402.13236},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

Experience

Apr 2026 - Present04/2026 - Now: Research Intern, Shanda AI Research Tokyo, Japan
Sep 2023 - Present09/2023 - Now: Ph.D. in Communication EngineeringCommE, National Taiwan University
Jan 202301/2023: M.S. in Computer ScienceCSIE, National Taiwan University
Jun 202006/2020: B.S. in Computer ScienceCSIE, Taiwan Tech

Selected Honors & Services

Research and Academic Honors

2026: NTU Mr. Wen Tzu-Hsiang Memorial Scholarship, one of only 9 recipients in NTU
2025: Best Student Paper nominee, The IEEE Autom. Speech Recognit. and Underst. Workshop
2025: Best Paper Award, The 37th Conf. on Comput. Linguist. and Speech Processing
2024-2025: Student Travel Grants, Google (2024) & ACLCLP in Taiwan (2025)
2024-2025: CTCI Foundation Bursary Award & Research Scholarship for Overseas Students,
2021: Ranked 3rd/42 submissions worldwide on the LA track of ASVspoof 2021 challenge
2020-2026: Kwong Tung Community Outstanding Student Scholarship, awarded six years.
2018-2020: Certificate of Achievement, Taiwan Tech (Top 5% of students; three semesters)
2017: 3rd Prize of SZIIT Academic Award , awarded by SZIIT (Top 20% of students)
2016: National Encouragement Scholarship, awarded by the Ministry of Education

International Event Organizer & Reviewer

2025: Co-Organizer, Responsible Speech & Audio Generative AI, Special Session at IEEE ASRU 2025
2024: Technical Committee, Codec-SUPERB Challenge, Special Session at IEEE SLT 2024
2023 - Now: Reviewer: AAAI, ICML, NeurIPS, ACL, EMNLP, COLM, ICASSP, ICME, INTERSPEECH, ASRU, SLT, COLING