Developed an AI system that identifies subtle facial micro-expressions during mock interviews, comparing machine detection accuracy against trained psychologists to explore perception gaps.

构建 AI 微表情识别系统，在模拟面试场景中捕捉细微面部信号，对比机器与专业心理学家的判断差异，探索感知盲区。

Applied NLP to extract and visualize trade patterns from historical merchant correspondence along the Silk Road, mapping commodity flows and cultural exchange across dynasties.

运用 NLP 从丝路商人往来书信中提取贸易模式，可视化历代商品流动与文化交流路径。

Leveraging AI-powered comparative analysis to reconstruct phonological patterns of an ancient language family by identifying systematic sound correspondences across modern descendant languages.

利用 AI 驱动的比较分析方法，通过识别现代后裔语言间的系统音变对应规律，重建古语族的音韵系统。

Transforms musical compositions into procedurally generated 3D visual stories, mapping harmony and rhythm to dynamic scene elements with real-time WebGL rendering.

将音乐作品转化为程序化生成的 3D 视觉故事，将和声与节奏映射为动态场景元素，实时 WebGL 渲染。

Built a system that generates short stories following genre conventions, with a multi-dimensional evaluation framework assessing narrative structure, character depth, and stylistic coherence.

构建短篇小说生成系统，遵循类型文学范式创作，并从叙事结构、人物深度、风格一致性等维度进行多维质量评估。

Developed an AI-powered framework analyzing how color usage in advertisements carries different cultural meanings across East Asian and Western markets, scoring brand-color alignment.

构建 AI 驱动的分析框架，研究广告中色彩运用在东亚与西方市场的文化含义差异，量化品牌-色彩适配度。