Information Discrepancy in Strategic Learning

Authors: Yahav Bechavod, Chara Podimata, Steven Wu, Juba Ziani

ICML 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We complement our theoretical analysis with experiments on real-world datasets.
Researcher Affiliation Academia 1School of Computer Science and Engineering, The Hebrew University. 2School of Engineering and Applied Sciences, Harvard University. 3School of Computer Science, Carnegie Mellon University. 4School of Industrial and Systems Engineering, Georgia Institute of Technology.
Pseudocode Yes Algorithm 1 Principal-Agent Interaction Protocol
Open Source Code Yes Our code is available in the supplementary.
Open Datasets Yes Available at https://archive.ics.uci.edu/ ml/datasets/default+of+credit+card+clients & https://archive.ics.uci.edu/ml/datasets/ adult.
Dataset Splits No The paper describes data preprocessing steps, group creation, and how projection matrices are obtained (e.g., "running SVD on the points inside of G1, G2"), but it does not provide explicit details about train/validation/test dataset splits with percentages or counts.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used to run the experiments.
Software Dependencies No The paper does not specify any particular software dependencies with version numbers (e.g., programming languages, libraries, frameworks) used for the implementation of the work.
Experiment Setup Yes In both cases, we ran ERM in order to identify w and we assumed that costs are A1 = A2 = Id d. [...] To obtain the projection matrices Π1, Π2, we ran SVD on the points inside of G1, G2. [...] Vg,5 correspond to the matrix having as columns the eigenvectors corresponding to the 5 top eigenvalues and zeroed out all other d 5 columns.