Kuan-Hao Huang

【一、個人Portfolio】

·       學校名稱：Texas A&M University (德州農工大學)

·       實驗室名稱：Frontier Language AI Research Lab

·       研究領域：自然語言處理 (NLP)、大型語言模型 (LLMs)、視覺語言模型 (VLMs)

·       職稱：Assistant Professor (助理教授)

·       電子郵件：khhuang@tamu.edu

·       個人網站：https://khhuang.me/

·       實驗室網站：https://khhuang.me/group.html

·       與臺大教授合作經驗：
資訊系林軒田教授，目前透過千里馬計畫host一位林教授的博士生。
電機系孫紹華助理教授：雙方實驗室有研究合作並已有論文產出。

【二、研究實習計畫】

·       研究內容簡介：開發語言AI Agent，教導其理解人類概念、使用人類語言及進行邏輯推理。目前核心方向包括：

1.     LLMs與VLMs的穩健性（Robustness）： 探討導致模型行為不可預測的挑戰（如對抗性輸入、跨領域不一致性），提升其在真實應用的穩定性。

2.     多模態推理(Multimodal Reasoning)：整合文本、影像與影片資訊，達成對跨模態概念的深度理解與推理。

3.     多語系理解(Multilingual Understanding)：研究模型如何習得跨語言的共享知識，以支持更公平、全球化的語言理解環境。

·       期待學生教育程度：大三（含）以上大學生、碩士生、博士生。

·       期待學生技能與特質：

o  對研究具有強烈熱忱、具備自信，且願意投入至少一年時間完成專題。

o  修習過機器學習(ML)及自然語言處理(NLP)相關課程。

o  具備半年以上AI相關研究經驗或大學專題生經驗。

o  未來有申請國外博士學位之打算。

o  (加分項目)具備以英文進行研究討論與口頭報告的能力。

·       實習行政資訊：

o  可接受期長：暑期

o  接受名額：2名。

o  簽證文件辦理：由PI透過學校國際處申請，辦理時長約 1-2 個月。

o  相關費用：PI負擔簽證服務費用(DS-2019)；學生需自行尋覓住宿（無補助）。

o  學分提供：否。

‍

I. Personal Portfolio

• University: Texas A&M University
• Laboratory: Frontier Language AI Research Lab
• Research Areas: Natural Language Processing (NLP), Large Language Models (LLMs), Vision-Language Models (VLMs)
• Position: Assistant Professor
• Email: khhuang@tamu.edu
• Personal Website: https://khhuang.me/
• Lab Website: https://khhuang.me/group.html
• Collaboration with NTU Faculty:

o  Prof. Hsuan-Tien Lin (Department of ComputerScience and Information Engineering): Currently hosting one of Prof. Lin’sPh.D. students through the NSTC Graduate Students Study Abroad Program.

o  Prof. Shao-Hua Sun (Department of ElectricalEngineering): The two labs have ongoing research collaboration and have alreadyproduced joint publications.

II. Research Internship Program

• Research Overview: The lab focuses on developing language AI agents capable of understanding human concepts, using natural language, and performing logical reasoning. Current core research directions include:

1.    Robustness of LLMs and VLMs:
Investigating challenges that lead to unpredictable model behavior (e.g.,adversarial inputs, cross-domain inconsistencies), with the goal of improvingreliability in real-world applications.

2.    Multimodal Reasoning:
Integrating text, images, and video to enable deep understanding and reasoningacross modalities.

3.    Multilingual Understanding:
Studying how models acquire shared knowledge across languages to support a moreequitable and globally inclusive language understanding environment.

• Target Student Level: Undergraduate students (junior year and     above), master’s students, and Ph.D. students
• Expected Skills and Qualifications:

o  Strong passion for research, confidence, andwillingness to commit at least one year to a research project

o  Completion of coursework in Machine Learning (ML)and Natural Language Processing (NLP)

o  At least six months of experience in AI-relatedresearch or undergraduate research projects

o  Intention to pursue a Ph.D. degree abroad in thefuture

o  (Preferred)Ability to conduct research discussions and oral presentations in English

• Administrative Information:

o  Available Duration: Summer

o  Number of Positions: 2

o  Visa Processing: The PI will apply through the university’s Office of International Affairs; processingtime is approximately 1–2 months

o  Costs: The PIwill cover the visa processing fee (DS-2019); students are responsible forarranging their own accommodation (no subsidy provided)

o  Academic Credit: Not offered