Grant: Alliance Grants

This project will apply computer science and data science in novel and effective fashions to public safety problems managed by the Saskatchewan Ministry of Corrections and Policing (""the Ministry""). The core activity of the Ministry is to oversee Saskatchewan's correctional system and facilities, and to provide guidance and support to Saskatchewan's municipal police services. Specifically, the goals of this project are to (A) improve monitoring of the use of the inmate telephone system (""Telmate"") by increasing the speed and decreasing the cost of transcribing inmate calls, (B) improve the monitoring of the same system by detecting anomalous patterns of usage, (C) increase the speed and decrease the cost of anonymizing Ministry data to allow analysis without endangering the privacy of those depicted in the data, (D) find a consistent, transparent way of performing risk assessment on inmates, and (E) estimating the trustworthiness of computerized agents. The potential outcomes, besides the research results themselves, are new policies and procedures for the Ministry. As examples: (A) more convenient transcription of telephone calls may allow security intelligence officers to change their procedures for analyzing security threats by allocating more time to analysis and less to listening, (B) detailed analysis of telephone call log data may lead to policies affecting the permissible frequency, duration, and times of inmate calls, (C) improved anonymization techniques may facilitate procedures for more data analysis projects, (D) an automated risk assessment tool may lead to new policies and procedures for inmate classification, and (E) improved understanding of trust in automated systems may contribute to the future formulation of new policies for automated vehicles. The primary benefit will be to aid the Ministry of Corrections and Policing in several initiatives related to public safety in Saskatchewan. The project is intended to increase public and inmate safety, obtained from detecting illegal behaviours, facilitating the analysis of confidential data, predicting good and bad behaviours in correctional facilities, and reducing the errors made by automated systems.