Fairness Through Computationally-Bounded Awareness

Authors: Michael Kim, Omer Reingold, Guy Rothblum

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We propose a new notion of fairness called metric multifairness and show how to achieve this notion in our setting. [...] At present, the results are theoretical, but we hope this work can open the door to empirical studies across diverse domains, especially since one of the strengths of the framework is its generality. We view testing the empirical performance of metric multifairness with various choices of metric d and collection C as an exciting direction for future research.
Researcher Affiliation Academia Michael P. Kim Stanford University mpk@cs.stanford.edu Omer Reingold Stanford University reingold@stanford.edu Guy N. Rothblum Weizmann Institute of Science rothblum@alum.mit.edu
Pseudocode Yes Algorithm 1: Switching Subgradient Descent Let > 0, T 2 N, and C 2X X . Initialize w0 2 F = [ B, B]n; W = ; For k = 1, . . . , T: If 9S 2 C such that ˆRS(wk) > 4 /5: // some constraint violated Sk any S 2 C such that ˆRS(wk) > 4 /5 wk+1 wk M 2 r RSk(wk) /* step according to constraint project onto F if necessary */ Else: // no violations found W W [ {wk} // update set of feasible iterates wk+1 wk GM r L(wk) /* step according to objective project onto F if necessary */ Output w = 1 |W | P w2W w // output average of feasible iterates
Open Source Code No The paper does not contain any statement about releasing source code or provide links to a code repository.
Open Datasets No The paper defines abstract data distributions ('Let D denote the data distribution over individuals and labels supported on X { 1, 1}') but does not specify any particular publicly available or open dataset used for training or evaluation.
Dataset Splits No The paper is theoretical and does not describe empirical experiments with dataset splits for training, validation, or testing.
Hardware Specification No The paper is theoretical and does not describe experimental procedures that would require specific hardware. Therefore, no hardware specifications are provided.
Software Dependencies No The paper is theoretical and describes algorithms and proofs, but does not mention specific software dependencies with version numbers for implementation.
Experiment Setup No The paper is theoretical and does not describe empirical experiments or their setup, thus no specific hyperparameter values or training configurations are provided.