Knowledge tracking method based on multidimensional cognitive space tomography decoupling and related equipment

By employing a knowledge tracking method that decouples knowledge points into cognitive attribute category probabilities and latent semantic vectors through multidimensional cognitive space layering, and combining sequence interaction coding with closed-loop state tracking units, the method addresses the shortcomings of existing methods in terms of interpretability and specificity, thereby improving the accuracy of knowledge tracking and prediction.

CN122388922APending Publication Date: 2026-07-14SOUTH CHINA NORMAL UNIV
4 Cites 0 Cited by

Patent Information

Application Number
CN202610511900.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-17
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing knowledge tracing methods lack in-depth mining of the potential cognitive attributes of knowledge points, resulting in insufficient interpretability and specificity, especially in cold start scenarios where prediction performance degrades.

Method used

A knowledge tracking method based on multidimensional cognitive space layer decoupling is adopted. The knowledge points are decoupled into cognitive attribute category probabilities and latent semantic vectors through multi-view feature extraction and fusion units and Gaussian mixture autoencoders. The method is then combined with sequence interaction coding and closed-loop state tracking units for dynamic updating and prediction.

Benefits of technology

It improves the interpretability and relevance of knowledge tracking, enhances the accuracy of prediction results, and is particularly effective in cold start scenarios by quickly generating dynamic feature representations that adapt to the current task.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a knowledge tracking method based on multidimensional cognitive space tomography decoupling and related equipment, and can be applied to the technical field of intelligent education. According to the method, the historical interaction sequence is input into the knowledge tracking model based on multi-dimensional cognitive space chromatography decoupling, feature vector extraction and fusion are carried out in the aspects of semantics, structures and statistics through the multi-view feature extraction and fusion unit in the model to obtain the target fusion features, and then the multi-dimensional attributes of the knowledge points can be described; then, knowledge points in the historical interaction sequence are decoupled into a cognitive attribute category probability and a potential semantic vector according to the target fusion feature vector by using a Gaussian mixture auto-encoder, and then potential cognitive dimensions are deeply mined; and then a sequence interactive coding and closed-loop state tracking unit is used for performing knowledge dynamic updating and closed-loop tracking according to the cognitive attribute category probability and the potential semantic vector, so that the interpretability and pertinence of knowledge tracking are effectively improved, and the accuracy of a knowledge tracking prediction effect is further improved.
Need to check novelty before this filing date? Find Prior Art