A simplified computational memory model from information processing.

Area of Science:

• Computational Neuroscience
• Cognitive Science
• Information Processing Theory

Background:

• Understanding memory mechanisms is crucial for cognitive science and artificial intelligence.
• Existing models often lack a unified information processing framework.
• Biological and graph theories offer potential for modeling neural structures.

Purpose of the Study:

• To propose a novel computational model for memory based on information processing principles.
• To simulate key memory functions such as memorization and strengthening.
• To provide a theoretical framework for understanding memory at a functional level.

Main Methods:

• Development of a bi-modular hierarchical functional memory network named SMIRN.
• Definition of meta-memory using biology and graph theories.
• Creation of intra-modular and inter-modular networks with a specific modeling algorithm.
• Introduction of a polynomial retrieval algorithm.

Main Results:

• The SMIRN model successfully simulates memory phenomena including memorization and strengthening.
• Theoretical analysis confirms the model's consistency with information processing views of memory.
• Simulation results validate the functional capabilities of the proposed network.

Conclusions:

• The SMIRN model offers a viable computational approach to understanding memory functions.
• This information processing framework provides insights into neural network dynamics.
• The model's bi-modular hierarchical structure is effective for simulating memory processes.