With the support of the legally-grounded methodology of situation testing, we tackle the problems of discrimination discovery and prevention from a dataset of historical decisions by adopting a variant of k-NN classification. A tuple is labeled as discriminated if we can observe a significant difference of treatment among its neighbors belonging to a protected-by-law group and its neighbors not belonging to it. Discrimination discovery boils down to extracting a classification model from the labeled tuples. Discrimination prevention is tackled by changing the decision value for tuples labeled as discriminated before training a classifier. The approach of this paper overcomes legal weaknesses and technical limitations of existing proposals.
In this paper we study the one commodity pickup-and-delivery traveling salesman problem with restricted depot (1-PDTSP-RD), which is a generalization of the classical traveling salesman problem (TSP). We first introduce a polynomial size integer programming formulation for the problem and then study the feasibility issue which is shown to be N P-complete by itself. In particular, we prove sufficient conditions for the feasibility of the problem and provide a polynomial algorithm to find a feasible solution. We also develop a bound on the cost of the 1-PDTSP-RD solution in terms of the cost of the TSP solution. Based on this bound, we provide a heuristic algorithm to solve the 1PDTSP-RD. Extensive numerical experiments are performed to evaluate the efficiency of both the exact and approximation algorithms.
Discrimination discovery from data consists of designing data mining methods for the actual discovery of discriminatory situations and practices hidden in a large amount of historical decision records. Approaches based on classification rule mining consider items at a flat concept level, with no exploitation of background knowledge on the hierarchical and inter-relational structure of domains. On the other hand, ontologies are a widespread and ever increasing means for expressing such a knowledge. In this paper, we propose a framework for discrimination discovery from ontologies, where contexts of prima-facie evidence of discrimination are summarized in the form of generalized classification rules at different levels of abstraction. Throughout the paper, we adopt a motivating and intriguing case study based on discriminatory tariffs applied by the U. S. Harmonized Tariff Schedules on imported goods
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.