Research

Representative Works

The following works best describe who I am as a researcher. For a list of complete works, see my CV.

Note: * indicates co-first authors

Consumer Choice and Digital Marketplaces

1. Learning Contextual Effects in Ordered Choice Sets: An Application to Online Marketplaces

with Patrick Ding. Working Paper (2026).

On online shopping and booking platforms, products are displayed as ranked lists, and platforms monetize position through sponsored placements. We study the counterfactual demand gains from sponsored placement by estimating contextual effects in large ordered choice sets with a computationally efficient EM algorithm for multinomial probit models. Using hotel search data from Expedia, we find patterns consistent with limited consumer attention and substantial heterogeneity in the value of sponsored spots, implying that platforms can capture more profit and provide better service through individualized pricing and recommendation structures.

2. On Sinkhorn’s Algorithm and Choice Modeling

with Alfred Galichon, Wenzhi Gao, and Johan Ugander. Operations Research (2026).

We connect matrix balancing and choice modeling, two topics that are almost 100 years old, and use the connections to obtain new insights on the convergence of Sinkhorn’s algorithm and on how the connectivity structure of choice data governs how efficiently preferences can be learned.

3. Handling Sparse Non-negative Data in Business and Economics

with Agostino Capponi. Under Review (2026).

Although Poisson pseudo maximum likelihood is commonly recommended for modeling data with non-negative dependent variables, such as sales and expenditure, there can be better estimators depending on the sparsity and heteroskedasticity of data. We propose a systematic framework that informs empirical researchers on such choices.

Causal Inference and Econometric Methods

1. Triply Robust Panel Estimators

with Susan Athey, Guido Imbens, and Davide Viviano. Journal of Applied Econometrics (2026). [Python Package]

We propose an estimator of causal effects in panel data that combines weighted two-way fixed effect regression with a low-rank factor, with weights based on similarity measures between units and between time periods, instead of regression-based weights.

2. Semiparametric Estimation of Treatment Effects in Observational Studies with Heterogeneous Partial Interference

with Ruoxuan Xiong*, Jizhou Liu*, and Guido Imbens. Journal of Business and Economic Statistics (2026).

We study augmented inverse propensity weighting (AIPW) estimators for causal effects under partial interference that accommodate heterogeneous interactions among units, relevant to settings such as spillovers of marketing campaigns and interventions in networks.

3. Distributionally Robust Instrumental Variables Estimation

with Yongchan Kwon. Under Review (2024).

We propose a distributionally robust version of the classical IV estimation method for inferring causal effects, motivated by common concerns in practice such as instrument validity and generalization across heterogeneous markets.

Operations Research and Machine Learning

1. Optimal Diagonal Preconditioning

with Wenzhi Gao*, Oliver Hinder, Yinyu Ye, and Zhengyuan Zhou. Operations Research (2025).

We provide SDP and interior point methods for finding the optimal diagonal preconditioners of a matrix.

2. Scalable Approximate Optimal Diagonal Preconditioning

with Wenzhi Gao, Madeleine Udell, and Yinyu Ye. Computational Optimization and Applications (2026).

We propose scalable methods to compute approximate optimal diagonal preconditioners applicable to sparse matrices of size up to 10^7.

3. Inferring Dynamic Networks from Marginals with Iterative Proportional Fitting

with Serina Chang*, Frederic Koehler*, Jure Leskovec, and Johan Ugander. 41st International Conference on Machine Learning (ICML) (2024).

We propose the ``biproportional Poisson’’ model, which provides statistical foundations for a widely used message passing algorithm to infer network traffic from marginal information.

4. Structured Lasso for convex nonparametric least squares: An application to Swedish electricity distribution networks

with Zhiqiang Liao. Major Revision at European Journal of Operational Research (2025).

We propose a structured Lasso method for variable selection in convex nonparametric regression problems.