The Relative Value of Prediction in Algorithmic Decision Making

Authors: Juan Carlos Perdomo

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our main results are theoretical in nature. We identify simple, sharp conditions determining the relative value of prediction visa-vis expanding access, within several statistical models that are popular amongst quantitative social scientists.
Researcher Affiliation Academia 1Harvard University, Center for Research on Computation and Society, Cambridge, MA. Correspondence to: Juan Carlos Perdomo <jcperdomo@g.harvard.edu>.
Pseudocode No No pseudocode or structured algorithm blocks are presented in the paper.
Open Source Code No The paper does not contain any statements about releasing source code for the described methodology or provide links to a code repository.
Open Datasets No The paper uses numerical simulations with chosen parameter values (e.g., 'we compute the cost benefit ratios with α = r2 = .01 and µ = 1, β = 10 for the linear case. For the probit case, we set b = .1 and α = r2 = 1e 3.'), not empirical datasets. Therefore, no information about public dataset access for training is provided.
Dataset Splits No The paper is theoretical and does not involve empirical model training or validation using dataset splits.
Hardware Specification No The paper is theoretical and uses numerical simulations without specifying any particular hardware used for computations.
Software Dependencies No The paper mentions 'The simulation code is included in the submission' in Appendix C but does not specify any software dependencies with version numbers for reproducibility.
Experiment Setup Yes For the visuaizations in Figure 1 and Figure 2, we compute the prediction access ratios numerically. Using our closed form expressions regarding the value functions V lin and V pr from Proposition 3.4 and Proposition 4.3 we compute the cost benefit ratios with α = r2 = .01 and µ = 1, β = 10 for the linear case. For the probit case, we set b = .1 and α = r2 = 1e 3.