Self-awareness and User-awareness Through Theory of Mind and Empathy

In the near future, intelligent software agents and robots have to take on more complex interactive tasks, such as being partners in offices or being companions for elderly people. Context-awareness, i.e. the ability to collect decision-relevant information from environment, is the first prerequisite for systems intended to have such a high degree of human interaction. In a previous project, we researched on currently used approaches to context-aware systems and their critical comparison. Moreover, we developed a cognitive software agent based on the LIDA (Learning Intelligent Distribution Agent) architecture that can show context-aware behaviors in a computer-simulated environment.

In this project, we intend to go one step further and try to develop user-awareness and self-awareness through Theory of Mind (ToM) and Empathy in intelligent software agents or robots that have to take on complex interactive tasks. While context-awareness is based on "perception" of the environment, self-awareness is based in particular on registration of important "inner" states that can influence decisions. Time components, Theory of Mind and Empathy are aspects of self-awareness and user-awareness that have to be taken into account in this project.

The project sets out with an interdisciplinary overview of different approaches to the above-mentioned aspects of self-awareness and user-awareness, including neuroscientific, sociological, psychological and philosophical perspectives as well as a critical overview of existing computational models of self-awareness and user-awareness. In a second step, we will attempt to design, implement and evaluate a computational model of self-awareness and user-awareness on one of the candidate cognitive architectures identified.

Research staff

• Robert Trappl
• Paolo Petta
• Soheil Khosravipour

Sponsor