AI-BASED VIRTUAL AVATAR FOR POLICE INTERVIEW TRAINING

The proposed project aims to revolutionise the way Child Protective Services and Law enforcement personnel are trained to conduct interviews of children. The idea is to build an avatar assisting the training and giving feedback on questions asked with help of novel explainable and interpretable artificial intelligence (AI) methods.
Master

The sexual and physical abuse of children is a significant societal problem in need of attention. Norway’s Child Protective Services received 57,988 allegations of child maltreatment in 2019 alone. An estimated 118 million children in Europe every year are victims of abuse, and 850 children under age 15 die as a result, according to the World Health Organization. The objective in this project is to protect vulnerable children from abuse, facilitate prosecution of offenders, and ensure that innocent adults are not accused of criminal acts. There is therefore a need for systems that help interviewer training to equip police with the skills to conduct high-quality interviews. To support this important task, we propose to research a training program that utilizes different system components and multimodal data from the field of artificial intelligence such as chatbots, generation of visual content, text-to-speech, and speech-to-text. The goal of combining all these different technologies and datatypes is to create an immersive and interactive child avatar that responds in a realistic way, to help to support the training of police interviewers, but can also produce synthetic data of interview situations that can be used to solve different problems in the same domain. This project has just started, and there are a large number of of open challenges like:

  • language models
  • AI-based chatbots-text-to-speech and speech-to-text
  • realistic AI-based visual avatars
  • how to detect online live abuse

We collaborate with social sciences at OsloMet, Child Protection Services (barnevernet), and several international partners.

Want to be part of our research group and improve the current systems? Contact us, and let’s see if we can find a thesis topic that matches your interests!

Goal

The overall objective is to build an avatar-based system to help improving interviewing skills of child-welfare and law-enforcement professionals - in the end helping to protect and assist sexually and physically abused children.

Learning outcome

interdisciplinary research
machine learning / AI
building systems

Qualifications

Hard working, motivated
Some knowledge of machine learning might be good
Interested in learning (the rest can be learned during the thesis work)

Supervisors

  • Pål Halvorsen
  • Michael Riegler

Collaboration partners

  • ForzaSys AS
  • Child Protective Services (CPS) and Law enforcement, Norway
  • Griffith University, Australia
  • University of Cambridge, UK

References

dl.acm.org/doi/10.1145/3463944.3469269
phillipi.github.io/pix2pix/
github.com/iperov/DeepFaceLab
[https://www.youtube.com/watch?timecontinue=30&v=2RaqHBClUQ](https://www.youtube.com/watch?timecontinue=30&v=2RaqHBClUQ)
github.com/albertpumarola/GANimation
www.youtube.com/watch?v=AmUC4m6w1wo

Contact person