An artificial neural network approach for the language learning model

Area of Science:

• Computational Linguistics
• Artificial Intelligence
• Machine Learning

Background:

• Language-based differential models are crucial for understanding learning processes.
• Developing accurate numerical solutions for these models is computationally challenging.
• Existing methods may lack efficiency or precision in capturing learning dynamics.

Purpose of the Study:

• To present numerical solutions for a language-based differential model using artificial intelligence.
• To implement and validate a scale conjugate gradient neural network (SCJGNN) procedure.
• To classify language learning into unknown, familiar, and mastered stages.

Main Methods:

• Utilized an artificial intelligence (AI) procedure based on a scale conjugate gradient neural network (SCJGNN).
• Employed the Adam scheme to minimize mean square error for dataset generalization.
• Configured the SCJGNN with a log-sigmoid activation function, 12 neurons, and specific layer structures, processing data in training (75%), validation (13%), and testing (12%) ratios.

Main Results:

• Achieved high accuracy in numerical solutions, with absolute errors ranging from 10^-06 to 10^-08 for all learning classes.
• Demonstrated perfect model performance through regression analysis for each learning stage.
• Validated the dependability of the SCJGNN approach using histogram and function fitness analyses.

Conclusions:

• The AI-based SCJGNN provides a robust and accurate method for solving language-based differential models.
• The model effectively distinguishes between unknown, familiar, and mastered language learning states.
• The SCJGNN approach shows significant potential for applications in computational linguistics and educational technology.