What Matheta is NOT
Matheta is not an AI agent. It is not an expert system, learning algorithm, demonstration of machine learning, or knowledge representation system.
What Matheta is
Matheta is a method that replicates the structure and functioning of animal nervous systems by modeling networks of neurons that detect the external world and produce actions in it. Like living organisms, actions are controlled by stimuli, drives, affects, classical conditioning, and learning based on consequences.
What is demonstrated on this website
Using a Roomba robotic base and an Arduino running a program simulating a dozen-neuron nervous system, we demonstrate numerous behavioral and learning phenomena including: stimulus-response reflexes, goal oriented behaviors, classical (Pavlovian) conditioning, operant conditioning in response to rewards and punishments, secondary reinforcement, and a simple affect response (e.g., fear).
What has been realized but is not yet demonstrated on this site
Prototypical affects as described by Silvan Tomkins: that is, simple examples of basic emotional responses shown by humans and other species (e.g., fear, contentment, interest). These affects are generated by abstract stimulating conditions (i.e., a particular level/rate of input stimuli), and produce internal, communicative, and behavioral responses, and can evolve through associative learning. Thus, they constitute the basis for learned emotions.
A complex nervous system architecture that recreates fundamental characteristics of animal behavior: Examples include the conflict between getting food and trying to avoid being consumed by other animals; minimizing energy expenditure in the pursuit of goals; seeking/doing that which is “pleasurable” and avoiding that which is “aversive.”
Multiple additional learning phenomena: including four forms of operant conditioning, conditioned emotional responses, extinction, pattern detection, discrimination and generalization.
What is possible with development
Matheta’s present capabilities are the product of a representation of a very simple “limbic system,” that is, one replicating the affective/emotional basis of behavior underlying most animal (and much human) behavior. It has virtually no “cognitive” capability. The addition of a full compliment of basic affects, additional drives (e.g., attachment), as well as the cognitive capacity of an added “cortex” would, eventually, allow replication of increasingly complex and realistic behavioral capabilities.
With Matheta you can:
- Replicate the behavioral and learning capabilities of organisms
- Provide mechanical/electrical systems with functional affect/emotions
- Create independently acting, evaluating, and learning systems that pursue user-defined goals