Who Leads and Who Follows in Strategic Classification?

Authors: Tijana Zrnic, Eric Mazumdar, Shankar Sastry, Michael Jordan

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We illustrate this empirically. In Figure 1 we generate non-strategic data according to y ~ Bern(p) and x0|y ~ N(4y - 2, 1) and plot the difference in risk between the two equilibria for the decision-maker and the agents, for varying B and p.
Researcher Affiliation Academia Tijana Zrnic University of California, Berkeley tijana.zrnic@berkeley.edu Eric Mazumdar California Institute of Technology mazumdar@caltech.edu S. Shankar Sastry University of California, Berkeley sastry@eecs.berkeley.edu Michael I. Jordan University of California, Berkeley jordan@cs.berkeley.edu
Pseudocode No The paper includes mathematical equations for updates, such as "φt+1 = ΠΘ(φt − ηt d/δ L( µt, φt + δut)ut)", but these are not presented as structured pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any explicit statement or link indicating the availability of open-source code for the methodology described.
Open Datasets No In Figure 1 we generate non-strategic data according to y ~ Bern(p) and x0|y ~ N(4y - 2, 1) and plot the difference in risk between the two equilibria for the decision-maker and the agents, for varying B and p. The paper describes data generation process rather than using a pre-existing public dataset.
Dataset Splits No The paper presents theoretical models and illustrates results with generated data in Figure 1, but it does not specify training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory, cloud resources) used for running the experiments.
Software Dependencies No The paper does not specify any software dependencies or their version numbers (e.g., programming languages, libraries, frameworks, solvers).
Experiment Setup No The paper describes the mathematical setups for linear and logistic regression and parameters for data generation (B and p in Figure 1), but does not provide specific experimental setup details such as hyperparameters (e.g., learning rate, batch size, optimizer) or system-level training settings.