Robots that collaborate with humans to form team plans
People, including doctors, nurses, and military personnel, often learn their roles in complex organizations through a training and apprenticeship process.
This learning through apprenticeship is made possible because humans have an innate ability to infer the mental state of others and to learn the implicit constraints and preferences of other team members through observation and interaction. This research draws insight from well-established human cognitive models and translates these cognitive models into computational models that robots use to infer our cognitive state and thereby collaborate with us better. The key challenge is to design computational models that support efficient inference, so that the machine can collaborate with us in real-time.