Speech AI Research center

徵人啟示

• 博士班學生：『人工智慧語音研發中心』研究助理，前瞻領域＋產學共創，穩定＆額外獎助學金

• 台語語料校正人員：教育部台語語料庫計畫（111-115）

• 計畫專任研究助理&工程師：臺灣智慧醫療創新加值推動計畫 - 以多模態大數據建構健康長壽與退化性疾病預測分析平台 (113-117)

vision (願景)

long-Term: Universal Human-Machine Interface (Conversational AI) for Intelligent Entities like Humanoids or Autonomation Systems

• In the near future, humans will likely communicate directly with intelligent entities such as humanoids or autonomation systems through natural spoken language, instructing them to collaborate seamlessly. Consequently, next-generation computers and robots must evolve to not only interpret human language accurately but also possess a profound understanding of world dynamics. Additionally, these systems should be capable of processing multi-modal information—integrating cues from speech, gestures, and environment—while using various tools, including programming and automation capabilities, to execute human commands effectively.

長期：通用人機界面（會話式 AI）應用於人形機器或自動化系統的智能實體

• 在不久的將來，人類將能夠通過自然的語言直接與人形機器或自動化系統等智能實體進行交流，指導它們協同工作。因此，下一代電腦和機器人不僅需要準確地理解人類語言，還必須具備對現實世界動態的深刻理解。此外，這些系統應能處理多模態信息——整合語音、手勢和環境中的線索，並利用各種工具（包括程式設計和自動化功能）來高效執行人類指令。

• Short-Term: Generative AI & Multi-Modal Foundation Model (see, listen, speak, read, write)

We aim to develop robots that can independently learn human language and understand the world by solely watching television. This knowledge will be applied in various fields, such as Smart Medicine, Healthcare, Auto Manufacturing, and Language Tutoring.

• 短期：多模態生成式AI (能看，能聽，能說，能讀，能寫)

我們將開發多模態生成式AI，同時處理與生成多種媒體資訊（包括語音、文字、圖片與視訊等等），建立大型多模態基石模型。並將這些模型應用於各個領域，例如機器人、智慧醫療、健康護理、自動化製造和語言教學等等任務。

• Short-Term: Multi-Modal Smart Medicine

• 短期：多模態智慧醫療

• Short-Term: Multi-Modal Language Learning System

• 短期：多模態本土語言學習服務平台

Research Areas (研究方向)

• Speech Recognition/Synthesis/Translation/Conversion/Enhancement

• Speaker/Language/Emotion Recognition

• Large Language Models, Multi-Modal (speech, text, image, video) Foundation Model, Intelligent Robots

• Embedding System, In-Memory/Storage Computing

• Computer Vision/Intelligent Manufacturing

• Natural Language Processing/Smart Medicine

• Blockchain, IoT Security, Authentication, Cryptology

• Low Earth Orbit Satellite Tracking

Highlights (亮點)

• Taibun (Taiwanese Hokkien) & Hakka Large Language Models（台客語大語言模型）

Working Horses - High Performance Computing (計算資源)

Laboratory (電資大樓8E04，中心實驗室)

Contact Us

• Director

Yuan-Fu Liao (廖元甫)

Full Professor, 

Institute of Artificial Intelligence Innovation, Industry-Academia Innovation School, National Yang Ming Chiao Tung University

Email: yfliao (at) nycu.edu.tw,

Web.: https://sites.google.com/nycu.edu.tw/sarc/members/prof-yuan-fu-liao

Address: EF-375, No. 1001, Daxue Rd. East Dist., HsinChu, 300093, Taiwan

Telephone: +886-3-5712121 ext. 58530

• Secretary

Email: sarc@nycu.edu.tw

Address: Room 8E04, Microelectronics and Information Research Center, No. 1001, Daxue Rd. East Dist., HsinChu, 300093, Taiwan

Telephone: +886-3-5712121 ext. 54554, 54555

• Address

Room 8E04, Microelectronics and Information Research Center, 1001 Ta Hsueh Road, Hsinchu, Taiwan

新竹市大學路1001號電子資訊大樓8E04室