Showing 1–2 of 2 results for author: Boxer, K
-
Auditing Predictive Models for Intersectional Biases
Authors:
Kate S. Boxer,
Edward McFowland III,
Daniel B. Neill
Abstract:
Predictive models that satisfy group fairness criteria in aggregate for members of a protected class, but do not guarantee subgroup fairness, could produce biased predictions for individuals at the intersection of two or more protected classes. To address this risk, we propose Conditional Bias Scan (CBS), a flexible auditing framework for detecting intersectional biases in classification models. C…
▽ More
Predictive models that satisfy group fairness criteria in aggregate for members of a protected class, but do not guarantee subgroup fairness, could produce biased predictions for individuals at the intersection of two or more protected classes. To address this risk, we propose Conditional Bias Scan (CBS), a flexible auditing framework for detecting intersectional biases in classification models. CBS identifies the subgroup for which there is the most significant bias against the protected class, as compared to the equivalent subgroup in the non-protected class, and can incorporate multiple commonly used fairness definitions for both probabilistic and binarized predictions. We show that this methodology can detect previously unidentified intersectional and contextual biases in the COMPAS pre-trial risk assessment tool and has higher bias detection power compared to similar methods that audit for subgroup fairness.
△ Less
Submitted 22 June, 2023;
originally announced June 2023.
-
The Case for MUSIC: A Programmable IoT Framework for Mobile Urban Sensing Applications
Authors:
Shiva R. Iyer,
Soumie Kumar,
Kate Boxer,
Fatima Zarinni,
Lakshminarayanan Subramanian
Abstract:
This vision paper presents the case for MUSIC, a programmable framework for building distributed mobile IoT applications for urban sensing. The Mobile Urban Sensing, Inference and Control (MUSIC) framework is contextualized for scenarios where a distributed collection of static or mobile sensors collectively achieve an urban sensing task. The MUSIC platform is designed for urban-centric sensing ap…
▽ More
This vision paper presents the case for MUSIC, a programmable framework for building distributed mobile IoT applications for urban sensing. The Mobile Urban Sensing, Inference and Control (MUSIC) framework is contextualized for scenarios where a distributed collection of static or mobile sensors collectively achieve an urban sensing task. The MUSIC platform is designed for urban-centric sensing applications such as location sensing on mobile phones for road traffic monitoring, air quality sensing and urban quality monitoring using remote cameras. This programmable system, at a high level, consists of several small sensors placed throughout a city on mobile vehicles and a centralized controller that makes decisions on sensing in order to achieve certain well-defined objectives such as improving spatial coverage of sensing and detection of hotspots. The system is programmable in that our framework allows one to create custom smart systems by writing custom control logic for sensing. Our contributions are two-fold -- a backend software stack to enable centralized control of distributed devices and programmability, and algorithms for intelligent control in the presence of practical power and network constraints. We briefly present three different urban sensing applications built on top of the MUSIC stack.
△ Less
Submitted 28 September, 2018;
originally announced September 2018.
