Morteza Dehghani
University of Southern California, CA·Computer Science
Police traffic stops are among the most common encounters between officers and the public. They can improve public safety, or they can escalate into violence.
- Fields
Computer Science
Social Science
Psychology
- Topics
artificial intelligence
machine learning
video language models
- Grants
- 1 known grant, through Sep 2028
What they work on
- Develop video language models, AI that jointly interprets video and speech, to produce accurate and explainable summaries of officer-driver communication in body-worn camera footage of traffic stops.
- Incorporate the viewpoints of retired officers and community members with varied police experiences into the models through personalized reinforcement learning from human feedback and multi-task learning, so they perform well for all stakeholder groups.
- Build a secure, anonymized research corpus of publicly released body-worn camera footage and a community platform for collaborative video annotation and shared-task evaluation.
More about the research
How an officer and a driver talk to each other matters for everyone's safety, for enforcing the law, and for public trust. Body-worn cameras record many of these encounters, but the footage is far too large for people to review by hand.
Why it matters
Police departments and local governments across the country could use these tools to review body-worn camera footage at lower cost. That could improve transparency and accountability and help refine best practices for safer traffic stops. Officers, drivers and communities stand to benefit if the tools help reduce escalation and build public trust.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Narges Norouzi
University of California-Berkeley, CA·Education Research
Model students' problem-solving strategies and algorithmic thinking in computer science courses by integrating fine-grained process data to trace competencies across levels of abstraction.
Cong Ma
University of Chicago, IL·Mathematics & Statistics·Early-career
Develop robust ranking algorithms for pairwise comparison data that tolerate ill-conditioned sampling of which pairs are compared and relax parametric modeling assumptions.
Ravi Mangal
Colorado State University-Fort Collins, CO·Computer Science
Design quantitative metrics for evaluating code generated by large language models using symbolic execution, which explores program paths with symbolic inputs, and Bayesian program analyses that reason probabilistically about program behavior.
The school
University of Southern California: 47 labs on PI Pages, $1.3B a year on research, 882 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.