Our world is changing incredibly! With the aim of advanced artificial intelligence, autonomous robots are entering not only to our workplace but also our homes. Due to this fact, robots should be designed to have social abilities. It is believed that one of the most important challenges that the areas of Artificial Intelligence (AI) and Robotics will face, is how to make robots more “real” and more “social”. Social interaction capabilities enable agents and robots to interact with humans and with other agents in a natural and tangible way. My main research focuses on the issues and methodologies for teaching or reprogramming social agents that can simulate human-like behaviors such as power, persuasion, and trust to be transparent and act naturally.
Mojgan Hashemian
PhD Student
Research
Interests
- Human-Computer Interaction
- Affective Computing
- Intelligent Games
- Cognitive Science
Research Projects
Social power is one of the most pervasive concepts in human societies due to its function as a social heuristic for decision making. Even though power has been previously explored as a social heuristic for agent behavior there are still significant research gaps regarding social intelligence for intelligent virtual agent believability in both inter-agent and humanagent interactions. In this project, we analyzed this gap from Emotional Perspective in highlighting its role in Decision making process.
In this project we propose a non-intrusive approach to recognize mood states, using emotional features observed in face. The proposed approach has been designed in two different methodologies, non-inductive and inductive methods. Both methods are tested and evaluated by three different experiments, and it is proved that they are promising.