BIIC LAB: 人本訊號運算研究室

26/07/2026

☕【 校友分享｜Alumni 】

六月茶會很開心邀請厲害的BIICer : 校友松霖 回到實驗室，和大家分享一路走來的求學與職涯經驗✧･ﾟ

從前往英國攻讀博士，到近期順利加入國際科技公司，松霖分享了申請準備、研究歷程與職涯選擇中的實用建議。

謝謝松霖回來陪大家聊聊，讓大家對UK PhD和未來職涯有了更多想像，期待這些經驗能成為 BIICers 前往下一站的小小助力 .ᐟ.ᐟ

So glad to have our amazing alumnus Song-Lin back at BIIC Lab for our June tea party to share his academic and career journey ✧･ﾟ
From pursuing a PhD in the UK to recently joining an international tech giant, Song-Lin dropped so many valuable tips on application prep, research life, and career choices. Thanks for coming back to chat with us and giving everyone a clearer picture of studying in the UK. Hope these insights help inspire our BIICers as they gear up for their next steps .ᐟ.ᐟ

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

25/07/2026

🏆【 實驗空間｜SPARK 】

跟著我們一起走進 BIIC Lab 的三個研究空間：
綜三館的 SP•ARK、資電館實驗室，以及位於清大山上的創新育成中心。

三個空間各自承載不同的研究任務，也陪伴 BIICers 度過討論、實驗與交流的日常。
其中，最具代表性的 SP•ARK，更常被大家稱為「清大最美實驗室」✧･ﾟ

李祈均老師一直很重視研究空間的美感，希望實驗室不只好用，也能讓創意自然發生。
SP•ARK 結合「Signal Processing and AI at Work」與 ARK 的意象，
以創意孵化器為概念，打造出能促進跨域互動的 Creative Incubator。

如今 SPARK 獲得 London Design Awards 肯定，
很開心這份對空間的想像受到認可，也讓 BIIC Lab 的研究日常被更多人看見 .ᐟ.ᐟ

獲獎頁面：
https://thelondondesignawards.com/winner.php?page=3&competition_id=4&category=34&rank=15&review=

Take a quick tour of BIIC Lab across our three spaces: SP•ARK in Building 3, our lab in the EECS Building, and the Innovation Incubation Center up on the hill.

Each space serves a unique purpose for our daily discussions and experiments. Our most iconic spot, SP•ARK, is even known as the prettiest lab at NTHU ✧･ﾟ

Professor Jeremy Lee has always prioritized aesthetic space design, aiming to create an environment where creativity flows naturally. Standing for "Signal Processing and AI at Work," SP•ARK was designed as a Creative Incubator to boost cross-disciplinary interaction.

We are super excited that SP•ARK just won a London Design Award! It’s amazing to see our vision for research spaces recognized on an international stage .ᐟ.ᐟ

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

24/07/2026

🎤 【 受邀演講｜Invited Talk 】

李祈均老師受邀參與 2026 TWSIAM 年會，
本次演講主題為「Advancing AI for Disordered Speech: Assessment, Accessibility, and Articulation」

老師以 3A Framework 為主軸，從 Assessment、Accessibility 到 Articulation，
分享如何重新思考語音 AI 在語音障礙評估、輔助溝通與發音修復中的角色。
透過近期研究成果，老師也說明隱私保護、語音轉換與人本 AI 技術如何讓更多不同型態的聲音，在數位時代被聽見、被理解，也被更好地支持。

偷偷預告老師目前 9 月已經敲定兩場演講，
演講王 Jeremy 的行程持續更新中，鎖定我們的最新動態，一起 follow 老師的最新演講消息 ✧･ﾟ

Professor Chi-Chun (Jeremy) Lee recently delivered an invited talk at the 2026 TWSIAM Annual Meeting, diving into "Advancing AI for Disordered Speech: Assessment, Accessibility, and Articulation"

Using the 3A Framework, Jeremy shared how speech AI is transforming disorder assessment, assistive communication, and articulation recovery. Highlighting our latest research, he explained how privacy tech, voice conversion, and human-centered AI help diverse voices be heard and understood in the digital era.

Quick heads-up, Jeremy already has two more talks lined up for September. Stay tuned for more updates ✧･ﾟ

活動網址：https://twsiam2026.math.fju.edu.tw/%E6%B4%BB%E5%8B%95%E8%8A%B1%E7%B5%AE/%E9%A0%98%E5%9F%9F%E8%AB%96%E5%A3%87_%E7%85%A7%E7%89%87

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

23/07/2026

🎓【 畢業祝福｜Graduation 】

恭喜實驗室的碩二同學們即將從 BIIC Lab 順利畢業啦 ✧･ﾟ
從一起討論拍照路線，到拍照當天活力滿滿的你們，
謝謝大家把青春的一段時間留在實驗室，也讓 BIIC Lab 多了好多溫暖的回憶。
祝你們帶著在實驗室累積的能力和勇氣，繼續奔向更大的世界!!

Huge congratulations to our second-year master's students on graduating from BIIC Lab ✧･ﾟ
From planning our photo routes to bringing so much energy on shoot day, thank you for spending a chapter of your youth with us and filling the lab with so many warm memories. Wishing you all the best as you take the skills and courage you've built here and head out to conquer the world!!

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

22/06/2026

老師新開的課程登上報導啦 🚀

這學期的「生成式AI實務與應用」，李祈均老師規劃推動，把 AI、大型語言模型與多模態應用帶進課堂，讓同學不只聽懂技術，也真的動手把想法做出來。

從課程設計、設備資源到助教支援，背後真的累積很多用心。電機系的同學們下學期歡迎修這堂課，一起把 AI 做成可以回應真實需求的作品！

📖攜手世界級企業 #清華電機 培育科技產業的領航者

在科技快速變革的時代，清華電機持續透過教育創新、前瞻研究、產學合作與國際鏈結，培養能引領產業發展、創造未來價值的科技人才。

從 #半導體、 #矽光子、 #高速通訊 到 #智慧科技應用，從實作導向課程到海外交流機會，感謝科技島（TechNice）專文介紹清華電機近年發展方向，以及人才培育與研究創新的成果。

歡迎閱讀完整報導👇
https://www.technice.com.tw/issues/semicon/226177/

22/06/2026

🎤 【 受邀演講｜Invited Talk 】

2026 年 5 月 29 日，李祈均老師受邀參與 BrainHack School 2026 產學交流活動。
不同於研究成果發表，這次演講以個人經歷與職涯選擇為主軸，老師分享自己如何在有趣且具挑戰性的道路上持續探索，談選擇、挑戰及面對挫折時的思考。

在演講中，李祈均老師分享自己如何從一次次選擇與挫折中，逐步走向研究與教職之路。從高中赴美、大學接觸電機與訊號處理，到博士班挑戰與回台建立研究團隊，每一段經歷都指向同一個核心：選擇有趣但困難的事，並努力把它做好。
老師也提醒同學，真正能打開可能性的，往往不是最安全的路，而是願意持續學習與承擔挑戰的心態。

BrainHack School 2026 透過課程、交流與產學對話，讓學生有機會接觸不同領域的研究者與產業觀點。期待這場分享能鼓勵更多年輕學子，在探索未來方向時，勇敢選擇自己真正感興趣也願意投入挑戰的道路。

近期演講影片
｜5/14 臨床聽診AI技術開發與驗證：https://www.youtube.com/watch?v=nFlyziXAH0A
｜5/29 My Personal and Professional Journey: Enjoy with Competitiveness：https://www.youtube.com/watch?v=IDmoyfQOD2k&list=PLwvUhR_TwpXsquq9h96FkB6-7XlS96FlV&index=3

▰▱▰▱▰▱ BRAINHACK SCHOOL ▰▱▰▱▰▱

On May 29, 2026, Prof. Chi-Chun Lee was invited to join the industry-academia exchange session at BrainHack School 2026.
Instead of focusing on research results, he shared his personal and professional journey, reflecting on choices, challenges, setbacks, and growth.

Prof. Lee spoke about his path from studying abroad in high school, discovering electrical engineering and signal processing, to pursuing a Ph.D. and later returning to Taiwan to build his research team.
Through these experiences, he encouraged students to choose paths that are both interesting and challenging, and to keep learning even when facing difficulties.

BrainHack School 2026 connects students with researchers and industry perspectives through courses, exchanges, and dialogue.
We hope this talk inspires more young students to explore their future paths with courage, curiosity, and the willingness to take on meaningful challenges.

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

04/06/2026

📝【 論文錄取｜Congratulations 】

恭喜 BIIC Lab 今年共有 9 篇論文被 Interspeech2026 接受！
也恭喜實驗室校友婉軒、博後 Orchid，以及博士生欣妤、星航、敬涵，碩班生家妤、傑翔、煒恆、盈瑄，大家的努力與成果，將在全球 speech AI 研究社群中分享交流！能在同一年有 9 篇論文入選，真的非常值得紀念，也為大家感到驕傲✨

🎙️ INTERSPEECH 是全球語音科學與語音技術領域最具代表性的國際頂尖會議之一。會議涵蓋語音辨識、語音生成、語音情緒、多模態互動與人機溝通等前沿研究，是全球 speech AI 研究者交流最新成果的重要平台。

再次恭喜各位作者！祝大家在 Sydney 順利分享研究成果，也在會議之餘好好享受澳洲的陽光與城市風景 ✈️

📝⋆｡˚ Our Interspeech 2026 Accepted Papers ˚｡⋆ 📝

Paper 01｜家妤 Chia-Yu Lee

Chia-Yu Lee, Huang-Cheng Chou, Tzu-Quan Lin, Yuanchao Li, Ya-Tse Wu, Shrikanth Narayanan, Chi-Chun Lee*
“Layer-wise Task Vector Merging: Leveraging ASR and SER Task Vectors for Enhanced Speech Emotion Representation”

Paper 02｜欣妤 Xin-Yu Chen

Xin-Yu Chen, Jing-Tong Tzeng, Carlos Busso, Chi-Chun Lee*
“Formant-Guided Speech Repair for Enhanced Comprehension of Dysarthric Speech”

Paper 03｜傑翔 Jie-Shiang Yang

Jie-Shiang Yang, Ya-Tse Wu, Chi-Chun Lee*
“Pitch-Injected Residual Adapter for Tonal Language in Neural Audio Codec”

Paper 04｜星航 Hsing-Hang Chou

Hsing-Hang Chou, Bo-Hao Su, Krishna Somandepalli, Chi-Chun Lee*
“Hidden Priors in Speech LLMs: Speaker Identity Shapes Emotional Perception”

Paper 05｜盈瑄 Ying-Hsuan Huang

Ying-Hsuan Huang, Woan-Shiuan Chien, Huan-Yu Chen, Chi-Chun Lee*
“Before the Turn: Investigating Motion Cues Preceding Speech in Dyadic Interaction”

Paper 06｜煒恆 Wei-Heng Huang

Wei-Heng Huang, Woan-Shiuan Chien, Chi-Chun Lee*
“Stress Detection Across Daily Activities: A Context-Aware Multimodal Framework with Trajectory and Ambient Speech”

Paper 07｜敬涵 Jing-Han Chen

Jing-Han Chen, Ya-Tse Wu, Bo-Hao Su, Xin-Yu Chen, Krishna Somandepalli, Chi-Chun Lee*
“CE-CoT: A Contrastive Empathetic Chain-of-Thought Training Strategy for Improving Emotion Consensus in Empathetic Speech LLMs”

Paper 08｜Orchid Chetia Phukan

Orchid Chetia Phukan, Girish, Mohd Mujtaba Akhtar, Chi-Chun Lee*
“Bridging the Age Gap: Towards Detecting Neural Audio Codec Synthesized Elderly Speech Deepfake”

Paper 09｜婉軒 Woan-Shiuan Chien

Woan-Shiuan Chien, Tomohiko Nakamura, Huan-Yu Chen, Satoru Fukuyama, Hitoshi Suda, Jun Ogata, Chi-Chun Lee*
“Two-Sided Fairness Transfer for Gender-Neutral Speech Emotion Recognition with Partially Observed Attributes”

── ✦ ── INTERSPEECH 2026 ── ✦ ──

Congratulations to BIIC Lab on having 9 papers accepted to Interspeech 2026!A big congratulations to our alumna Woan-Shiuan, postdoc Orchid, Ph.D. students Xin-Yu, Hsing-Hang, and Jing-Han, and master’s students Chia-Yu, Jie-Shiang, Wei-Heng, and Ying-Hsuan.Your hard work and research will be shared with the global speech AI community. Having 9 papers accepted in the same year is truly a milestone, and we are so proud of you all✨

🎙️ INTERSPEECH is one of the world’s leading international conferences in speech science and speech technology.
The conference covers cutting-edge research in speech recognition, speech generation, speech emotion, multimodal interaction, and human-machine communication, serving as a major platform for the global speech AI research community.

Congratulations again to all authors! Wishing you a successful presentation in Sydney, and a wonderful time enjoying Australia’s sunshine and city views ✈️

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

25/05/2026

📚【 論文交流｜AI 】

臺大人工智慧研究中心（AINTU）在 2026 年 5月 15 日主辦「AI 論文分享交流會 AI Research Exchange Workshop」，聚焦 AI 頂尖國際會議論文分享與跨域研究交流。活動邀請資深學者與頂會論文作者分享研究歷程、技術亮點與投稿經驗。透過講座、論文分享與海報交流，讓研究者能更快速掌握 AI 前沿趨勢，也開啟跨校、跨域合作的更多可能。本次 BIIC Lab 由博士生吳亞澤與陳敬涵出席參與交流。

亞澤分享他的論文 “ASR for Affective Speech: Investigating Impact of Emotion and Speech Generative Strategy”，探討情緒語音與生成式策略如何影響自動語音辨識效能。研究進一步提出目標式資料增強方法，協助 ASR 系統在真實情緒語音中降低字詞錯誤率，並提升對情緒語音的辨識穩定性。

敬涵則分享的 RE-LLM 研究，關注生成式 AI 在人機互動中的同理心表達。不同於僅依賴文字資訊的模型，RE-LLM 結合語音中的情緒細節，讓系統能更深入理解使用者的情緒狀態。研究結果顯示，這樣的設計能提升 emotional reaction 與 emotional exploration 等同理心表現，讓 AI 回應更貼近人。

透過兩篇研究分享，我們看見 AI 如何從「聽懂語音」進一步走向「理解情緒」。謝謝兩位準博士亞澤與敬涵的精彩分享，展現 BIIC Lab在前沿 AI 研究上的持續探索，推動更貼近人的智慧互動💡

活動網址：https://reurl.cc/bdo1Ey

─── ⋆⋅ PAPER ⋅⋆ ─── AI ─── ⋆⋅ EXCHANGE ⋅⋆ ───

The AI Research Exchange Workshop, hosted by the NTU Artificial Intelligence Research Center (AINTU), focused on top-conference AI papers and cross-disciplinary research exchange.
Through talks, paper sharing, and poster sessions, the event helped researchers connect with the latest AI trends and explore new collaboration opportunities. BIIC Lab was represented by Ph.D. students Ya-Tse Wu and Jing-Han Chen shared his work on RE-LLM, which aims to make speech-based LLM responses more empathetic.

Ya-Tse presented “ASR for Affective Speech,” exploring how emotional speech and generative strategies affect automatic speech recognition.
His work proposes targeted data augmentation to improve ASR performance on real emotional speech.

Jing-Han shared his work on RE-LLM, which focuses on empathetic responses in generative AI.
By integrating emotional cues from speech, RE-LLM helps AI better understand users’ emotional states and generate more human-centered responses.

Through these two research talks, we saw how AI can move from understanding speech to understanding emotion.
Thank you to our two Ph.D. candidates, Ya-Tse and Jing-Han, for the insightful sharing and for showcasing BIIC Lab’s continued exploration of more human-centered AI.

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

22/05/2026

🎤 【 受邀演講｜Invited Talk 】

2026 年 5月 14 日，李祈均老師受邀出席於南港展覽館舉辦的「在宅醫療照護跨域交流會」，與產官學研醫各界交流智慧醫療的最新發展，並分享 BIIC Lab 在智慧醫療與臨床聽診 AI 上的研究成果。

本次活動由國科會指導、國科會補助工程科技推展中心主辦，並結合 ATLIFE 2026 臺灣輔具暨長期照護大展。現場匯集政府、醫界、學界、研究機構與產業代表，共同討論智慧科技如何回應高齡照護與在宅醫療需求。

在宅醫療的核心，不只是把醫療服務帶回家中，更是讓照護能更貼近真實生活情境。我們也從 AI 聽診技術出發，思考如何讓臨床經驗、資料驗證與科技研發形成更完整的照護解方。

李祈均老師本次演講主題為「臨床聽診AI技術開發與驗證」。演講從聽診器在心音與肺音判讀中的重要性談起，也指出傳統聽診仍仰賴專業人員現場判讀，容易受到場域、經驗與主觀差異影響。從肺音分類、心音異常判讀，到與臺大新竹醫院、廣達合作進行臨床驗證，BIIC Lab 的研究不只追求技術表現，也重視真實醫療場域中的可用性。

未來，當電子聽診器與 AI 判讀結合，期望能以此連結臨床與日常，讓智慧醫療更自然地進入居家、偏鄉與長照場域，真正走進生活！

─── ⋆⋅ ⌁ ⋅⋆ ── AI × HEALTHCARE ── ⋆⋅ ⌁ ⋅⋆ ───

On May 14, 2026, Prof. Chi-Chun Lee was invited to speak at the “Interdisciplinary Forum on Home Healthcare” held at Taipei Nangang Exhibition Center, sharing BIIC Lab’s research on smart healthcare and clinical auscultation AI.

Guided by the National Science and Technology Council and organized by the Engineering Promotion Center, the forum brought together experts from government, industry, academia, research, and medicine to discuss the future of home healthcare and aging care.

Home healthcare is not only about moving medical services into the home, but also about making care more connected to real-life needs.
Through AI-based auscultation, BIIC Lab explores how clinical experience, data validation, and technology development can come together in practical care solutions.

Prof. Lee’s talk, “Development and Validation of Clinical Auscultation AI Technology,” introduced BIIC Lab’s work on respiratory sound classification, cardiac abnormality detection, and clinical validation with National Taiwan University Hospital Hsin-Chu Branch and Quanta.

By combining electronic stethoscopes with AI interpretation, we hope to bring smart healthcare more naturally into homes, rural communities, and long-term care settings.

─── ⋆⋅ ⌁ ⋅⋆ ── AI × HEALTHCARE ── ⋆⋅ ⌁ ⋅⋆ ───

｜活動影片：https://youtu.be/BaEc2Fy6Ncw

｜新聞報導：https://www.ctee.com.tw/news/20260518700655-431204

— Be Innovatively Creative — ılılı BIIC Lab @ NTHU EE

20/05/2026

辛苦實驗室博士生 陳欣妤 做的漂漂手牌！
還有碩班生 溫家馨 當助教協助課程🙌🏻
大家都好棒🫶🏻