Non-Discriminatory Machine Learning Through Convex Fairness Criteria
Authors: Naman Goel, Mohammad Yaghini, Boi Faltings
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experimental analysis demonstrates the success of the method on popular real datasets including Pro Publica s COMPAS dataset. and 7 Experimental Evaluation As mentioned earlier, the weighted sum of logs technique needs to be tuned for meeting a given p-rule requirement while ensuring as little drop in accuracy as possible. We evaluate three additional techniques that can used for this. ... We work with two datasets in our experimental evaluation. The first dataset is the Pro Publica s COMPAS dataset (Larson et al. 2016). ... The second dataset is the popular Adult Income dataset (UCI 1996). |
| Researcher Affiliation | Academia | Naman Goel, Mohammad Yaghini, Boi Faltings Artificial Intelligence Laboratory, Ecole Polytechnique F ed erale de Lausanne, Lausanne, Switzerland, 1015 {naman.goel, mohammad.yaghini, boi.faltings}@epfl.ch |
| Pseudocode | No | The paper describes the proposed technique in text and mathematical formulations but does not include any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper does not provide an explicit statement or link for the open-source code of the methodology described in this paper. |
| Open Datasets | Yes | The first dataset is the Pro Publica s COMPAS dataset (Larson et al. 2016). It contains data from Broward County, Florida, compiled by Pro Publica. ... The second dataset is the popular Adult Income dataset (UCI 1996). This dataset has 45, 222 instances and 14 attributes such as age, occupation, working hours per week and work class etc. ... References: Larson, J.; Mattu, S.; Kirchner, L.; and Angwin, J. 2016. https://github.com/propublica/compas-analysis, 2016. UCI. 1996. Adult income data set. url: https://archive. ics.uci.edu/ml/datasets/adult. |
| Dataset Splits | No | All the results presented in this paper are k-fold cross-validated. The paper mentions using k-fold cross-validation but does not specify the value of k or provide explicit train/validation/test split percentages or sample counts. |
| Hardware Specification | No | The paper does not provide any specific hardware details such as GPU or CPU models, memory specifications, or types of computing resources used for the experiments. |
| Software Dependencies | No | The paper mentions using a 'logistic regression classifier' and states 'Our implementation uses the same set of libraries and code for optimization and cross validation etc.', but it does not specify any particular software names with version numbers. |
| Experiment Setup | No | The paper discusses hyperparameters like δ and γ conceptually, but it does not provide specific numerical values for these or other experimental setup details such as learning rates, batch sizes, or optimizer configurations used in the experiments. |