2024.04 — Present
Nara Institute of Science and Technology (NAIST)
Ph.D. Candidate in Information Science
Human AI Interaction Laboratory · Advisor: Prof. Dr. Sakriani Sakti
Haotian Tan 谭皓天
Ph.D. Candidate in Information Science
Human AI Interaction Laboratory
Nara Institute of Science and Technology (NAIST), Japan
Haotian Tan is a Ph.D. candidate in Information Science at the Nara Institute of Science and Technology (NAIST), where he is a member of the Human AI Interaction Laboratory under the supervision of Prof. Dr. Sakriani Sakti.
His research focuses on simultaneous speech translation and real-time spoken language systems. In particular, he studies low-latency translation policies and system architectures for continuous speech, together with related topics in automatic speech recognition, speech synthesis, and speech language models.
Education
2023.04 — 2024.03
Japan Advanced Institute of Science and Technology (JAIST)
Ph.D. Student in Information Science
Advisor: Prof. Dr. Sakriani Sakti · Transferred to NAIST in April 2024.
2021.04 — 2023.03
Kanazawa University
M.Eng. in Information Engineering
Advisor: Prof. Dr. Junichi Akita
2015.09 — 2019.06
Guilin University of Electronic Technology
B.Eng. in Electronic Science and Technology
Guilin, China
Research Interests
Simultaneous Speech Translation
Low-latency translation of continuous and unsegmented speech, with an emphasis on quality–latency trade-offs and READ/WRITE decision policies.
Automatic Speech Recognition
Robust speech understanding and streaming ASR components for multilingual speech translation systems.
Speech Synthesis
Incremental TTS and speech-to-speech translation for natural real-time cross-lingual communication.
Speech & Language Models
Efficient integration of large language and speech models into practical real-time spoken interaction systems.
Publications
2026
Learning When to Speak: Human-Inspired Unit Sensing for Machine Simultaneous Interpretation
IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP).
2026
2025
NAIST Simultaneous Speech Translation System for IWSLT 2025
Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025), pp. 369–378.
2025
NAIST Offline Speech Translation System for IWSLT 2025
Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025), pp. 360–368.
[Paper]System paper
2024
Contrastive Feedback Mechanism for Simultaneous Speech Translation
Proceedings of INTERSPEECH 2024, pp. 852–856.
2024
Chunk Size Scheduling for Optimizing the Quality-Latency Trade-off in Simultaneous Speech Translation
27th Conference of the Oriental COCOSDA International Committee (O-COCOSDA 2024), pp. 1–6.
[Paper]Conference
2024
NAIST Simultaneous Speech Translation System for IWSLT 2024
Proceedings of the 21st International Conference on Spoken Language Translation (IWSLT 2024), pp. 170–182.
[Paper]System paper
Research Projects
Participation: April 2024 – March 2026
A Study on Multi-modal Automatic Simultaneous Interpretation System and Evaluation Method
JSPS KAKENHI · Grant-in-Aid for Scientific Research (S) · 21H05054
Within this project, Haotian led the design and implementation of an LLM-driven real-time speech-to-speech simultaneous interpretation prototype. His work integrated streaming speech translation, READ/WRITE decision policies, and incremental speech synthesis, with a focus on balancing translation quality and latency. He delivered a working research prototype for experimental evaluation and system demonstrations.