Estimating Performative Effects in Dynamical Systems: the advantage of sequential observations
Primary Area: societal considerations including fairness, safety, privacy
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: performativity, dynamical systems, causal inference, control theory
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Abstract: Regulators and academics are increasingly interested in the causal effect that algorithmic actions of a digital platform have on consumption, a quantity in the machine learning literature termed the performative effect. In this work, we first show how isolated (non-sequential) observations are not enough to identify the performative effect of interest in general, then we show how sequential observations overcome these limitations. The key novelty of our approach is to explicitly model the dynamics of consumption over time, viewing the platform as a controller acting on a dynamical system. From this dynamical systems perspective, we are able to show that exogenous variation in consumption and appropriately responsive algorithmic control actions are sufficient for identifying the performative effect of interest. Our results illustrate the fruitful interplay of control theory and causal inference, which we illustrate with examples from econometrics, macroeconomics, and machine learning.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 6087
Loading