Workshop on Machine Learning meets Econometrics (MLECON)

A synergy between machine learning (ML) and econometrics (ECON) has created a growing body of works in several directions including nonparametric instrumental variable regression, proximal causal learning, double machine learning, adaptive allocation in economic systems and program evaluation, among others. In particular, there is great interest in ML for econometrics. ML for economics offers the potential for better predictions and handling larger, multimodal data to help address substantive questions in economics and the social sciences. Non- and semi-parametric econometrics naturally interface with modern machine learning. On the other hand, incorporating economic modeling of microfoundations can also improve machine learning’s impact in societal systems and policy-relevant questions.

The MLECON workshop aims to serve as an interface for experts from both disciplines to meet and exchange ideas and for participants to present their works-in-progress that lies at this intersection. We welcome both extended abstract and technical papers on topics that shed light on the area, including (but not limited to) the following questions:

• When can we address unobserved confounding in policy evaluation? Conversely, how can we create predictive models that are robust to hidden confounders?

• Can model selection be used in identifying the best economic models from data?

• How can we use machine learning to improve generalized method of moments (GMM) and conversely, can moment selection improve machine learning?

• How do we model and empirically calibrate economic, social, and ethical concerns that arise from the use of machine learning models in the real world?

• How do we leverage machine learning models when the ultimate goal is economic performance such as revenue or welfare?

• How can we apply semi-parametric models in machine learning?

• How can we apply recent advances in offline reinforcement learning to policy evaluation and learning in economics?

• Can econometric techniques such as instrumental variables, difference-in-difference, and regression discontinuities be used to improve machine learning pipelines?

• How can machine learning improve computationally difficult traditional econometric estimation such as dynamic discrete choice?