Chonghuan Wang 「王崇焕」

I am now a fourth-year PhD candidate in Data Science Lab at MIT, advised by Prof. David Simchi-Levi. I am affiliated with the Laboratory for Information and Decision Systems, Center for Computational Science and Engineering and Department of Civil and Environmental Engineering, MIT. Prior to joining MIT, I received my bachelor's degree in information engineering from Shanghai Jiao Tong University in 2020 advised by Prof. Haiming Jin and Prof. Xinbing Wang.

I am selected as 2023-2024 Accenture Fellow!

My research focuses on leveraging statistical learning techniques to design efficient, reliable and sustainable field experiments in clinical trials, online platforms, healthcare, etc. My research goal is to enhance social good with the power of statistical learning, econometrics theory and experimental design. I am also broadly interested in operations research, machine learning, econometrics and their interplay.

Email: chwang9 [at] mit.edu

Multi-armed bandit has been well-known for its efficiency in online decision-making in terms of minimizing the loss of the participants' welfare during experiments (i.e., the regret). In clinical trials and many other scenarios, the statistical power of inferring the treatment effects (i.e., the gaps between the mean outcomes of different arms) is also crucial. In this paper, we investigate the trade-off between efficiency and statistical power by casting the multi-armed bandit experimental design into a minimax multi-objective optimization problem.

One of the primary objectives of classical experiments is to estimate the average treatment effect (ATE) to inform future decision-making. However, in healthcare and many other settings, treatment effects may be non-stationary, meaning that they can change over time, rendering the traditional experimental design inadequate and the classical static ATE uninformative. In this work, we address the problem of non-stationary experimental design under structured trends by considering two objectives: estimating the dynamic treatment effect and minimizing welfare loss within the experiment.

When launching a new product, historical sales data is often not available, leaving price as a crucial experimental instrument for sellers to gauge market response. When designing pricing experiments, there are three fundamental objectives: estimating the causal effect of price (i.e., price elasticity), maximizing the expected revenue through the experiment, and controlling the tail risk suffering from a very huge loss. In this paper, we reveal the relationship among such three objectives.

Motivated by the empirical evidence observed from the real-world dataset, we study context-based dynamic pricing with separable demand models, where the price and the context influence the demand separately. We investigate the statistical complexity of the problem under different structural properties.

Updated in Sept 2023. Thanks Jon Barron for the source code. Thanks my friend Minkai Xu for letting me know this great template.