Meritocratic Fairness for Cross-Population Selection
Authors: Michael Kearns, Aaron Roth, Zhiwei Steven Wu
ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We complement our theoretical results with empirical simulations which emphasize that both the utility and fairness guarantees of our algorithms are better in practice than our theorems promise. ... 7. Simulations We conclude by discussing some illustrative simulation results for FAIRTOP, along with comparisons to simpler algorithms without fairness guarantees. |
| Researcher Affiliation | Academia | Michael Kearns 1 Aaron Roth 1 Zhiwei Steven Wu 1 1University of Pennsylvania, Philadelphia, USA. |
| Pseudocode | Yes | Algorithm 1 NOISYTOP({y1, y2, . . . , yn}, α, k) ... Algorithm 2 FAIRTOP(X = {xij}, ε, δ, k, m) ... Algorithm 3 ABOVETHRE(X = {xij}, ε, δ, k, m) |
| Open Source Code | No | The paper does not include a statement about open-sourcing the code or a link to a code repository. |
| Open Datasets | No | The simulations were conducted on data in which the raw scores for each population i = 1, 2 were drawn from N(µi, 1) respectively, and the µi themselves were chosen randomly from N(0, 1). ... The paper uses synthetically generated data and does not refer to a publicly available dataset with concrete access information or citation. |
| Dataset Splits | No | The paper describes generating synthetic data for simulations but does not specify training, validation, or test dataset splits. |
| Hardware Specification | No | The paper does not specify any hardware details (e.g., GPU models, CPU types, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not specify any software names with version numbers for reproducibility. |
| Experiment Setup | Yes | The simulations were conducted on data in which the raw scores for each population i = 1, 2 were drawn from N(µi, 1) respectively, and the µi themselves were chosen randomly from N(0, 1). While we varied the population sizes, they were held in the fixed ratio m1/m2 = 2 and k = 0.1(m1 + m2) . ... Figure 1(a) shows the underlying scores computed by FAIRTOP ... Overlaid on this arc of underlying scores is a black plot illustrating sample post-noise scores when ε = 10. |