Aim: To propose a model, which could be used in a simulation or real deployment of a robotic system that operates in a social environment. This work could then influence the design of social robots operating in tourist areas (for assisting tourists), in museums (for guiding visitors) or in airports (for passengers’ assistance but also for reinforcing security). How can we bridge together artificial intelligence technologies, such as deep learning, knowledge representation and intelligent decision making, together with brain cognition studies?
Background: The human brain uses the rational artifacts of space and time in order to make sense of the world, and to facilitate interaction and communication with fellow humans, and the ambient environment. As robotic systems move from the industrial setting to service-based robots for humans in everyday life scenarios, it is important to understand how the human brain works, in order to design robots which are empathic, being able to integrate well to their operational environment. One of the most important aspects to study, in this context, is human memory. This project is about mimicking the way humans store information to their brain, associating this information with other memories and happenings in the past. For example, an image of some person eating an apple could associate this person with the fact that he likes fruits, particularly apples. In the future, if the same person expresses hunger, the robot might suggest eating an apple. Or when the robot recognizes an apple, it might remember this same person if not a stronger stimulus appears at its memory. Thus, this model can then be used to improve the social behavior of robots when interacting with humans.
The development of a model for a service robot which operates in a specific application area and combines sensory and memory/historical information to perform intelligent decision making in an empathetic way.
Type of work expected:
- Basic research in new forms of neural and/or deep learning architectures and structures/formations of knowledge representation.
- Sentiment analysis to understand the emotional state of humans and how to act accordingly.
- Which other sensors (beside camera) can be used to recognize humans and infer their state or needs?
- This project builds upon prior work on this topic, performed on a conceptual-only level.
This project will take in co-operation with Pervasive Systems and CYENS Centre of Excellence.