Compact internal representation of dynamic situations: neural network implementing the causality principle.

Area of Science:

• Computational Neuroscience
• Artificial Intelligence
• Robotics

Background:

• Navigating dynamic environments requires efficient internal representations of complex, time-evolving situations.
• Understanding how the brain forms compact internal representations (CIRs) is a key challenge in neuroscience.

Purpose of the Study:

• To propose an artificial neural network model capable of generating CIRs for dynamic environments with moving obstacles.
• To investigate the use of causality and concurrent processes for simplifying time-dependent situations into static patterns.

Main Methods:

• Developed a neural network exploiting the principle of causality within a mental world model.
• Implemented two concurrent processes: wavefront interaction with obstacles and a diffusion-like relaxation process.
• Demonstrated the generation of CIRs as single points in a multidimensional phase space.

Main Results:

• The network successfully creates CIRs of dynamic situations, reducing complexity to static patterns.
• CIRs can be unfolded for flexible, task-dependent path planning in real-time environments.
• The network supports "autonomous thinking" by evaluating mental situations without direct motor execution.

Conclusions:

• The proposed artificial neural network effectively models CIRs for dynamic environments using causality.
• This approach enables efficient path planning and offers potential for autonomous cognitive functions.
• Hypothesizes a neuronal mechanism for detecting spatio-temporal coincidences in dynamic environments.