Juliusz Tarnowski, 28 May 2020
Definition of a social robot
The robot is a embedded agent. Social robots are distinguished by the fact that their embodiment and behavior are designed with particular attention to collaboration in a group - in a robot or human community. The robot is an embedded agent. Social robots are distinguished by the fact, that their embodiment and behavior should allow them to collaborate with groups of other robots or humans. The social robot must be able to communicate in a way that is understandable and intuitive for others and it also should be able to recognize each other. In particular, the social robot must be able to engage in social interactions in form of: expressing emotions, creating relationships and learning from other group members (e.g. by imitating gestures). The robot should have an individual character, history and should be able to interpret the world in terms of its own experience (to make the relationship with it unique and not generic).
Leslie's model of theory of mind and Baron-Cohen's model of theory of mind and implications of these models to humanoid robots
According to Leslie, the world is naturally decomposed into three classes of events: one class for mechanical agency, one for actional agency, and one for attitudinal agency. He states that nature's evolution produced independent domain-specific modules to deal with each of these events. His Theory of Mind assumes that thoughts, beliefs, and desires are real and should be considered regardless of whether or not the circumstances involved are real. Because of that, it explains events (agent's actions) in terms of the intent and goals of agents.
Baron-Cohen’s model assumes two forms of information are available as input: self-propeled stimuli and direction and eye-like stimuli. The first input describes all visual, auditory, and tactile acutators with self-propelled motion. The second input describes all visual stimuli that are eye-like. Baron-Cohen's theory of mind (or his “mindreading system”), can be decomposed into four modules:
- Intentionality Detector describe the basic movements of approach and avoidance (“he wants the food”).
- The Eye Direction Detector detects the eye-like stimuli in visible field, decides whether agent is looking at him or something else and presents results in form of “agent sees me”.
- The Shared Attention Mechanism takes the information from 1st and 2nd and produces representations of the form “I see the girl”.
- The Theory of Mind Mechanism provides a way of representing mental states of agent and a mechanism for tying together knowledge of this states into a coherent whole, giving the full range of mental concepts.
The motivations that have driven the development of humanoid robots
- The basic motivation for developing anthropomorphic structures is the desire to have a machine that is adapted to work in a human environment. The humanoid shape allows the use of tools intended for humans. The workplace can still be used by the human beings because of the human-like movement of the robot.
- The human brain is mainly adapted to interact with other people. A person loses interest in still objects much faster than in a human face or human voice. Therefore, embedding the robot in a humanoid structure can be much more effective in terms of interaction compared to other possibilities. Such a robot can more effectively serve as an assistant and more efficiently convince, e.g. to buy a given product.
- Humankind has the desire to "create a being in their own image". In many areas, achieving competence at a human level is seen as a milestone.
- Humanoid can be a platform to better understand the functioning of the body and mind of man. When developing such a construction, you can encounter problems that our body solves intuitively, and which are very difficult to implement from scratch.
- Humanoid robots can help develop prosthesis or become an artificial body for people with disabilities. They can also be used for entertainment (games or sports based on artificial man control).