I have a broad interest in many areas. As a senior student, I have been searching for my own research field throughout these four years, and I have found two areas that I'm most interested in, natural language processing and causal inference.
You can find my GitHub Repo here. My Google Scholar page is this.
2022
PROJECTS
Synthetic Control Methods for Estimating the Effect of the Change of Leadership in Early-stage Startups [paper]
2021
My gap year at Sinovation Ventures.
2020
Conference Paper
Zhu, Chenyang; Abrahao, Bruno; Zhao, Anqi; and Paolo Parigi, "The Effect of User Interactions on Shaping Online Trust: Evidence from a Large-scale Experiment" (2020). PACIS 2020 Proceedings. 45. [link]
2019
PROJECTS
Tutorial Slides: Introduction to Chinese Natural Language Processing, Chenyang Zhu, 2019 [slides]
Otter, a Python deep learning framework [GitHub] A complete deep learning framework, with auto-gradient, computation graph and useful CNN/RNN structures, benchmarking TensorFlow and Pytorch. Noticeably, backpropagation and gradient updates are combined into one step with a single depth-first search on the computation graph to accelerate computation.
Measuring the roles of different channels for political news spread Chenyang Zhu, Wenqi Chen, and Lujia Bay, 2019 [paper] Created a partial differential and probabilistic model to show how society responds to news from different channels (e.g., state media, public media, or individuals).
Comparison of Algorithms for Noisy First Order Gradients in Stochastic Accelerated Optimization, Chenyang Zhu, 2019 [paper][GitHub] Compared algorithms solving optimization problems with noisy first-order gradient descents on convergence rate in terms of their bias and variance bounds. Strongly convex and non-strongly convex cases were computed to compare the methods for optimal convergence.
Trade-off Between Safety and Congestion: An Application in Traffic Reinforcement Learning, C. Zhu, W. Chen, A. Li, K. Yuan, Y. Zhang, 2019 [paper][GitHub] Developed models based on stochastic processes and multi-agent reinforcement Q-learning to optimize traffic around our school, with the aim to reduce student injuries caused by over-speeding cars and distracted deriving. The paper was delivered to school faculties and has raised heated discussion in students and faculties.
COURSE THESIS
Into Interpretable US-China Trade War, Chenyang Zhu, 2019, SUFE Machine Learning Course Thesis 中文水课论文;]
Cloud Detection at Poles with Bancroft Transformation, Chenyang Zhu and Ling Xie, 2019 paper, Github
2018
PROJECTS
AcNet: Accumulative Deep Learning Neural Network, Jiaye Teng and Chenyang Zhu, 2018 [draft] (This work remain unpublished or will not be updated from the current draft.)
Smart City Cooperation Project with Statistics Bureau, Shanghai Government C. Zhu , W. Chen, D. Liu, X. Chen, Z. Zhu. We interviewed firms from Hangzhou and benchmarked policies in Beijing and Shenzhen. We estimated the potential effect of new policies in shanghai transportation and potential revenue of new bus lines. We provided solutions to tackle parking and public transportation problems in a city of 30m population. Download our public report here (Report is in Chinese) 《政企共建智慧城市迈入多元化高速通道 释放科技活力须解码制度创新人才优先 》
COURSE THESIS
Successful SARIMA model with Information Theory and Cross Validation, Chenyang Zhu, 2018 paper
2017
PROJECTS
Pattern Recognition for Spam Review Based on Text Mining, Yutan Han, Chenyang Zhu, 2017. Scraped spam reviews from online shopping platform Taobao.com and created our own spam review dataset by sending out surveys to writers. Detected spam reviews at 80% accuracy with PE algorithm, after cutting sentences into words with Jieba.
Investigating the Value of New-round Supply Side Revolution to Leisure Agriculture: an Example of Fengxian Shanghai Y. Chen, Y. Wang, Y. Han, A. Li, W. Li, C. Zhu We won Second Prize Paper in Shanghai Zhixing Cup!