Open Access

Nonlinear Adaptive Optimization of Multi-Modal Learning Paths Using Graph Convolutional Networks and Reinforcement Learning for Intelligent Educational Systems

TongLI,

TongLI,

  • School of Information Engineering, Beijing Institute of Graphic CommunicationBeijing, China
  • School of Computer and Communication Engineering, University of Science and TechnologyBeijing, China
Search for this author on
Sciendo|Google Scholar
TongLI,
  
Mar 17, 2025
Nonlinear Adaptive Optimization of Multi-Modal Learning Paths Using Graph Convolutional Networks and Reinforcement Learning for Intelligent Educational Systems's Cover Image
About this article
Previous Article
Next Article
Cite
Download Cover
Published Online: Mar 17, 2025
Received: Oct 19, 2024
Accepted: Feb 04, 2025
DOI: https://doi.org/10.2478/amns-2025-0829
Keywords
© 2025 TongLI, published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1.

Infrastructure components of the designed system
Infrastructure components of the designed system

Figure 2.

Pseudo code of the “Multimodal Data Fusion and Temporal Modeling”Algorithm
Pseudo code of the “Multimodal Data Fusion and Temporal Modeling”Algorithm

Figure 3.

Recommendation Accuracy with primary metrics
Recommendation Accuracy with primary metrics

Figure 4.

The results of time series modeling using RMSE
The results of time series modeling using RMSE

Figure 5.

Diversity and Learning Outcome Analysis Result
Diversity and Learning Outcome Analysis Result

Figure 6.

Experimental Group vs. Control Group Analysis
Experimental Group vs. Control Group Analysis

Space Complexity of each module in the proposed framework

Module Space Complexity Optimization Suggestions
User and Resource Embeddings O(Nu·dg + Nm·dg) Reduce embedding dimensions dg.
Knowledge Graph Storage O(|Ekn|). Ekn

summarizes the time complexity of each module in the proposed model_

Module Time Complexity Optimization Suggestions
Multi-modal Fusion Module O(M·dm) Reduce the feature dimension dm.
Temporal Modeling (LSTM) O(T⋅dt2)\[\text{O}(\text{T}\cdot \text{d}_{\text{t}}^{2})\] Decrease the sequence length T.
Temporal Modeling (Attention) O(T2·dt) Apply sparse attention mechanisms.
Reinforcement Learning Module O(T·df) Enhance parallelization of training.

Performance of ablation experiments

Model Configuration Precision@5 Recall@5 Diversity
Content Only 0.632 0.511 0.581
Code Only 0.647 0.520 0.562
Video Only 0.624 0.495 0.534
Full Model 0.732 0.641 0.693

Analysis of the Temporal Modeling Module

Model Configuration Diversity CTR (Click-Through Rate) Learning Completion Rate
No RL 0.612 0.465 0.602
Full RL Module 0.693 0.514 0.681
Language:
English
Publication timeframe:
1 times per year
Journal Subjects:
Life Sciences, Life Sciences, other, Mathematics, Applied Mathematics, General Mathematics, Physics, Physics, other
Journal RSS Feed