Police Decision Making
Polis, drawing on some of the landmark work of Prof. Anders Ericsson, have developed and implemented curriculum using the concept of deliberate practice to assess and improve police decision-making. Tactical Decision Exercises (TDEs), are intended to provide structured, realistic exercises designed to improve core cognitive and perceptual skills essential for effective decision-making. TDEs are based on extensive findings by Ericsson and others that deliberate practice of relative cognitive and perceptual skills, when done with immediate, focused feedback, can produce significant improvements in the development of decision-making essential for mission-critical performance.
While police departments have been using training videos in various ways for decades, the presentation and discussion of videos is typically unstructured, subjective, and inconsistent. As a result, many video training exercises unhelpfully reinforce the inaccurate notion that officers'
decision-making processes involving the use of force and other critical situations is largely a matter of individual perception, and thus impossible to standardize or evaluate in systematic fashion. This misperception is captured in the time-worn adage popular among many police officers that
"if you weren't there, you can't say anything."
Polis will conduct a National Institute of Justice funded study to assess cognitive performance-based training to improve police decision making.
To learn more about the study click here.
Patrol Expert Data Analytics
Polis built a computer model that identifies when officers are thinking more or less like typical expert or novice peers. Structural Topic Modeling (STM), is a statistical approach to language that goes deeper and shows how the words associated with expertise are associated with an underlying concept or “topic.” We examined 7 topics which showed conceptual/cognitive differences between experts and novices.
Experts were distinguished from Novices by an emphasis on de-escalation, emphasizing the rendering of aid, having a positive finish to an encounter. Novices greatly emphasized using force as a relevant course of action and interestingly, were slightly more likely than experts to emphasize a duty to intercede during the inappropriate conduct of others.
Wearable Sensing and Human Performance
Polis Co-Founders Drs. Brian Lande and Jonathan Wender oversaw the use of devices called "sociometric badges" for measuring social performance in mission-critical environments as part of the Defense Advanced Research Projects Agency's (DARPA) SSIM program. The research Brian and Jonathan conducted in this area contributed to the development of the T3™ training model by helping to identify key features of successful social interaction.
Brian conducted further research with Aptima on the use of wearable sensors for measuring rapport, social influence, and interpersonal coordination during mission-critical tasks. Brian and his colleagues at Aptima and Sociometeric Solutions Inc. will be publishing a series of papers on the use of sociometers in this domain.
Brian also collaborated with Ben Waber at Sociometeric Solutions and Dr. Elizabeth Torres at Rutgers University to examine the use of wearable sensors to support the rehabilitation of patients with Traumatic Brain Injuries and comas (Disorders of Consciousness). This innovative work helps medical teams make sense of interpersonal movement by integrating data from wearable accelerometers, gyroscopes, and thermometers. The research by Brian and his colleagues has been covered by the mainstream media (BusinessWeek and the Fresno Bee) and is also in review by peer-reviewed medical journals. Their work promises to open up a new field of medical telemetry for people with brain injuries.
Computational Social Vision
Brian introduced a new sub-discipline of computer vision during his time at DARPA, which he calls "computational social vision." This field, developed in collaboration with Ajay Divarakan at the prestigious nonprofit research corporation, SRI-Sarnoff, looked into how humans "see" "social things." This research identified the observable features of positive social interactions, such as rapport and trust. We are excited to use this cutting-edge research in our T3™ training. This research has been published here:
Salter, D., Tamrakar, A., Siddiquie, B., Amer, M., Divakaran, A., Lande, B., and Mehri, D. (2015). The Tower Game Dataset: A Multimodal Dataset for Analyzing Social Interaction Predicates. International Conference on Affective Computing and Intelligent Interaction.
Video Analytics for Public Safety (VAPS)
Polis participated in the inaugural meeting of a new federal effort called VAPS. Our combined expertise in social science research and police operations is seen by VAPS leadership as valuable to the urgent project of developing objective standards for analyzing video footage of police-community interactions. The goal of VAPS is to create a team of technical stakeholders to jointly identify the critical R&D, resources, standards, and collaboration vehicles and infrastructure to support the creation of a robust public safety video analytics R&D ecosystem. VAPS activities will help to focus and link research, collaboration, and standards activities across stakeholders to support the development of a robust and interoperable public safety video analytics ecosystem that meets the nation's growing needs.
In addition to training and consulting, we provide social science research support to police and criminal justice agencies, the Defense Department, national security organizations, and to some of the nation's top corporations and research institutions.
Polis stands out in the consulting community for our ability to collaborate with practitioners in policing and the military, as well as with engineers and technology developers.