Yuhang Zhou

I'm a First-year Ph.D. candidate at the School of Computer Science and Technology, Fudan University, Shanghai, China. I am also conducting research at the Institute of FinTech at Fudan University and Shanghai Innovation Institute . I am advised by Prof. Guangnan Ye and Prof. Yixin Cao at FDU and SII, also received guidance from Prof. Xipeng Qiu. Previously, I obtained a Bachelor's degree in Internet of Things Engineering from Jiangnan University in 2022, advised by Prof. Ya Guo at Key Laboratory of Advanced Control in Light Industry Processes, Ministry of Education.

Email / Scholar / Github

Research: Scalable Evaluation and Evolution of LLMs, Fintech, AI for Art

I am very interested in the scalable evaluation and evolution of large language models, aiming to enable iterative evolution of LLMs and enable comprehensive evaluation in a more automated way. At the same time, I have also done some fintech work in the past.

Scalable Evaluation and Evolution of LLMs: Current evaluation of large models relies excessively on costly, inefficient human annotations and static benchmarks, struggling to address the challenges posed by dynamic environments, continuous learning requirements, and reliability demands in high-stakes scenarios—while also overlooking the potential of inexpensive yet effective signals from weak models and rules. My focus is on building an automated evaluation and evolution framework based on weak supervision, leveraging small models or rule-based systems to provide scalable, low-cost feedback that drives large models toward continuous self-improvement, cross-domain adaptability, and safe, controllable long-term evolution.
Fintech: Using technologies such as LLMs, graph models, and time-series models to solve problems in the financial field such as intelligent investment advisory, intelligent text generation, and risk control.
AI for Art: Current multimodal models are re-empowering artistic creation across various domains, including but not limited to visual arts, theater, and music. My research primarily focuses on how AI can collaborate with artists and enhance interactive artistic creation.

Publications

	Teach2Eval: An Indirect Evaluation Method for LLM by Judging How It Teaches Yuhang Zhou, Xutian Chen, Yixin Cao, Zhiheng Xi, Yu He, Siyu Tian, Xiang Liu, Jian Zhang, Chuanjun Ji, Guangnan Ye, Xipeng Qiu Under Review, 2025 arXiv
	Are Large Language Models Rational Investors? Yuhang Zhou, Yuchen Ni, Zhiheng Xi, Zhangyue Yin, Xiang Liu, Jian Zhang, Sen Liu, Yixin Cao, Xipeng Qiu, Guangnan Ye, Hongfeng Chai Findings of ACL, 2025 arXiv
	R3-NL2GQL: A Model Coordination and Knowledge Graph Alignment Approach for NL2GQL Yuhang Zhou, Yu He, Siyu Tian, Yuchen Ni, Zhangyue Yin, Xiang Liu, Chuanjun Ji, Sen Liu, Xipeng Qiu, Guangnan Ye, Hongfeng Chai Findings of EMNLP, 2024 arXiv
	SilverSight: A Multi-Task Chinese Financial Large Language Model Based on Adaptive Semantic Space Learning Yuhang Zhou, Zeping Li, Siyu Tian, Yuchen Ni, Sen Liu, Guangnan Ye, Hongfeng Chai CCL / Journal of Chinese Information Processing, 2024 arXiv
	Research on Improving the Performance of Tri-ellipsoid Multi component Optical Sensors Yuhang Zhou, Yuchen Ni, Wei Ge, Ya Guo Laser & Optoelectronics Progress, 2023 arXiv

Selected Projects

ChatReport: A Financial Research Report Generation Approach Based on Internet of Agents
Yuhang Zhou, Yuchen Ni, Yunhui Gan, Siyu Tian, Jiahui Zhang, Hongjie Xia, Yu Liu, Guangnan Ye
Using multi-agent collaboration technology based on LLMs to complete the writing of financial multimodal long texts (such as financial research reports, bond applications, etc.), using a combination of large and small models to extract causal relationships from past research report data, simulating multi role interactions to form a document writing team, and achieving low-cost, long text financial document writing. The project has been implemented and applied in companies such as Guotai Junan Securities, DataGrand, and Baidu.
National 1st Prize, the Graduate Financial Technology Innovation Competition, 2023
Top 10 Excellence Awards, the 2024 LLM Financial Application Innovation and Practice Competition, 2024

Ink Wash Horizons: Digital Light and Shadow
Yuhang Zhou*, Hoho Wang*
Based on a collaborative modeling approach, this work aims to build an AI system that simulates Mr. Ku Chan’s artistic expression by decomposing his characteristics into two components: cognitive thinking patterns and artistic style. The core technical pathway consists of two dimensions: on one hand, a large language model is trained on Mr. Ku Chan’s textual corpus—including books, interviews, and creative notes—to construct a model of his cognition and aesthetic decision-making, thereby simulating his creative logic and expressive style; on the other hand, an image generation model performs fine-grained representation learning on a dataset of his artworks to build a high-fidelity style generation engine. The two components are cascaded to achieve synergy: the language model outputs semantic descriptions and compositional logic aligned with the artist’s cognitive traits, which serve as conditional inputs to drive the image generation model in producing visually coherent works with both stylistic fidelity and semantic depth. The overall architecture forms a closed loop of “semantic generation → visual mapping,” realizing a digital twin that bridges simulated artistic thinking with authentic visual style reproduction. This approach demonstrates significant advantages in style controllability, interpretability of the creative process, and artistic consistency, offering a technically rigorous yet humanistically grounded pathway for AI-driven artistic creation.
Exhibited in Baotu Spring Scenic Area(趵突泉景区), Jinan City, Shandong Province.
Aired on CCTV-1's(央视一套) "Chinese Ink Art Gala".

Internship

New York University Shanghai(NYUSH) - Algorithm/Embedded System - Research Assistant/Teach Assistant (Sept. 2024 - Now)
Great Wall Fund - Financial quantification - Research Intern (Jul. 2023 - Nov. 2023)
Wind Inc. - Financial quantification - Research Intern (Feb. 2023 - Jul. 2023)
Huachuang Securities - Financial NLP tasks - Research Intern (Dec 2022 - Jan. 2023)

Invited Talks

Fintech: The Dual Journey of Financial Data and AI Models, Fudan University Graduate Student Association, Shanghai, Nov. 2024

Selected Awards

National 1st Prize, the Graduate Financial Technology Innovation Competition, 2023
First Class Academic Scholarship, Fudan University , 2022~2024
Outstanding Graduation Program(recommended for Provincial Outstanding), Jiangnan University, 2022
National 2nd Prize, The Blue Bridge Cup Software and Information Technology Competition, 2022
Huawei Excellent Scholarship, Jiangsu, 2022
Outstanding Youth League Cadres, Jiangnan University, 2022
Academic Scholarship, Jiangnan University, 2019~2022