Communications of the ACM
Sampling Near Neighbors in Search for Fairness
Credit: Getty Images
Similarity search is a fundamental algorithmic primitive, widely used in many computer science disciplines. Given a set of points S and a radius parameter r > 0, the r-near neighbor (r-NN) problem asks for a data structure that, given any query point q, returns a point p within distance at most r from q. In this paper, we study the r-NN problem in the light of individual fairness and providing equal opportunities: all points that are within distance r from the query should have the same probability to be returned. The problem is of special interest in high dimensions, where Locality Sensitive Hashing (LSH), the theoretically leading approach to similarity search, does not provide any fairness guarantee. In this work, we show that LSH-based algorithms can be made fair, without a significant loss in efficiency. We propose several efficient data structures for the exact and approximate variants of the fair NN problem. Our approach works more generally for sampling uniformly from a sub-collection of sets of a given collection and can be used in a few other applications. We also carried out an experimental evaluation that highlights the inherent unfairness of existing NN data structures.
1. Introduction
In recent years, following a growing concern about the fairness of algorithms and their bias toward a specific population or feature, there has been an increasing interest in building algorithms that achieve (appropriately defined) fairness.14 The goal is to remove, or at least minimize, unethical behavior such as discrimination and bias in algorithmic decision making, as nowadays, many important decisions, such as college admissions, offering home loans, or estimating the likelihood of recidivism, rely on machine learning algorithms. While algorithms are not inherently biased, nevertheless, they may create it by careless design, or by amplifying the already existing biases in the data.
There is no unique definition of fairness (see Hardt et al.18 and references therein), but different formulations that depend on the computational problem at hand, and on the ethical goals we aim for. Fairness goals are often defined in the political context of socio-technical systems and have to be seen in an interdisciplinary spectrum covering many fields outside computer science. In particular, researchers have studied both group fairness (also known as statistical fairness), where demographics of the population are preserved in the outcome,12 and individual fairness, where the goal is to treat individuals with similar conditions similarly.14 The latter concept of "equal opportunity" requires that people who can achieve a certain advantaged outcome, such as finishing a university degree, or paying back a loan, have an equal opportunity of being able to get access to it in the first place.
No entries found