Fair Forests: Regularized Tree Induction to Minimize Model Bias
By Edward Raff, Jared Sylvester, Steven Mills. 2017
76 citations recorded by the source corpus.
The potential lack of fairness in the outputs of machine learning algorithms has recently gained attention both within the research community as well as in society more broadly. Surprisingly, there is no prior work developing tree-induction algorithms for building fair decision trees or fair random forests. These methods have widespread popularity as they are one of the few to be simultaneously interpretable, non-linear, and easy-to-use. In this paper we develop, to our knowledge, the first technique for the induction of fair decision trees.We show that our "Fair Forest" retains the benefits of the tree-based approach, while providing both greater accuracy and fairness than other alternatives, for both "group fairness'' and "individual fairness.'' We also introduce new measures for fairness which are able to handle multinomial and continues attributes as well as regression problems, as opposed to binary attributes and labels only. Finally, we demonstrate a new, more robust evaluation procedure for algorithms that considers the dataset in its entirety rather than only a specific protected attribute.
Subfields: Artificial Intelligence, Safety Research
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Silvia Chiappa
- Serena Wang
- Luca Oneto
- Harikrishna Narasimhan
- Andrew Cotter
- Maya R. Gupta
- Julien Ferry
- Sébastien Gambs
In Safety Research
- Luca Oneto
- Silvia Chiappa
- Andrew Cotter
- Julien Ferry
- Serena Wang
- Harikrishna Narasimhan
- Sébastien Gambs
- Maya R. Gupta
Citations in the corpus
- Learning Fair Rule Lists (2019, methodology): “To compare with results found in FairForest [52], we also use the inconsistency metric (also defined in Section 2.1) as well as the delta , which is defined as follows. delta = accuracy − unf airness To build the Pareto front for each dataset, we use Algorithm 1.”
- Pairwise Fairness for Ranking and Regression (2019, background): “Raff et al. [2018] considered continuous protected features in their tree-growing criterion that addresses fairness.”
- Fairness in Machine Learning (2020, methodology): “More recently, regression [15, 22, 55, 156] has also started to be considered, as well as continuous sensitive attributes through discretization [103, 104, 149, 182].”