A Penalized Likelihood Method for Balancing Accuracy and Fairness in Predictive Policing

Abstract

Racial bias of predictive policing algorithms has been the focus of recent research and, in the case of Hawkes processes, feedback loops are possible where biased arrests are amplified through self-excitation, leading to hotspot formation and further arrests of minority populations. In this article we develop a penalized likelihood approach for introducing demographic parity into point process models of crime. In particular, we add a penalty term to the likelihood function that encourages the amount of police patrol received by each of several demographic groups to be proportional to the representation of that group in the total population. We apply our model to historical crime incident data in Indianapolis and measure the fairness and accuracy of the two approaches across several crime categories. We show that fairness can be introduced into point process models of crime so that patrol levels proportionally match demographics, though at a cost of reduced accuracy of the algorithms.

Learning to rank spatio-temporal event hotspots

“…While hotspot policing has been shown to yield crime rate reductions, there is the possibility of unwanted side effects of hotspot policing such as traffic stops that unfairly target minority populations, stop and frisk, and other police activities that have negative societal consequences. There has been some recent work on improving fairness of spatial crime forecasting algorithms (Wheeler 2019;Mohler et al 2018) where a fairness penalty is added to the optimization algorithm. Future research may focus on incorporating fairness into learning to rank models of crime, similar to methods that incorporate fairness into learning to rank for information retrieval (Zehlike and Castillo 2018).…”

Allocating Police Resources While Limiting Racial Inequality

“…In particular, it should reduce the number of minorities stopped and arrested by police. But that reduction comes with a cost of not equivocally targeting the highest crime locations in a hot spots policing strategy (Mohler et al, 2018).…”

“…Current critiques of predictive algorithms often state they ignore how historical processes shape the current data that is fed into the system, and as such likely perpetuate the current statusquo (Harcourt, 2007;Lum & Isaac, 2016). Predictive algorithms are not limited to a simple objective of improving forecast accuracy though -they can be amended to take into account additional objectives (Hardt, Price, & Srebro, 2016;Mohler et al, 2018). As opposed to saying predictive algorithms will hopelessly exacerbate racial disparities already prevalent in the criminal justice system, this is but one example of how they can be amended to attempt to reduce racial disparity.…”

“…If one can identify what unfair outcomes one wishes to avoid, it can be incorporated into the decision making process for how the information is used, or into the predictive algorithm itself (Corbett-Davies et al, 2017;Kleinberg et al, 2018). Deciding what is fair though is a difficult task -fairness along some particular continuum inevitably involves trade-offs with some other metric (Berk, 2016;Mohler et al, 2018;Ridgeway, 2017b).…”

“…Limiting police resources in certain areas no doubt creates an ethical dilemmaintentionally diverting resources from the highest crime areas could be discriminatory, as it deprives minority communities of needed protection from the police (Mohler et al, 2018).…”