Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Endowing Interpretability for Neural Cognitive Diagnosis by Efficient Kolmogorov-Arnold Networks

Shangshang Yang, Linrui Qin, Xiaoshan Yu

TL;DR

The paper tackles the interpretability gap in neural cognitive diagnosis models by introducing KA2NCD, which uses efficient Kolmogorov-Arnold networks to replace opaque MLP components or form a two-level KAN-based aggregation. It develops two deployment modes: KA2NCD-native that substitutes FC layers with KANs, and KA2NCD-e/KA2NCD-kan that processes embeddings with a two-level KAN architecture, combining lower-level and upper-level KANs to produce predictions. The authors implement training-time accelerations for KANs and validate KA2NCD across four real-world educational datasets, showing improved or competitive predictive performance and interpretability on par with traditional CDMs, with reasonable training costs. Overall, KA2NCD demonstrates that structured KAN-based architectures can provide transparent, interpretable cognitive diagnostics without sacrificing accuracy, enabling more trusted personalized education applications.

Abstract

In the realm of intelligent education, cognitive diagnosis plays a crucial role in subsequent recommendation tasks attributed to the revealed students' proficiency in knowledge concepts. Although neural network-based neural cognitive diagnosis models (CDMs) have exhibited significantly better performance than traditional models, neural cognitive diagnosis is criticized for the poor model interpretability due to the multi-layer perception (MLP) employed, even with the monotonicity assumption. Therefore, this paper proposes to empower the interpretability of neural cognitive diagnosis models through efficient kolmogorov-arnold networks (KANs), named KAN2CD, where KANs are designed to enhance interpretability in two manners. Specifically, in the first manner, KANs are directly used to replace the used MLPs in existing neural CDMs; while in the second manner, the student embedding, exercise embedding, and concept embedding are directly processed by several KANs, and then their outputs are further combined and learned in a unified KAN to get final predictions. To overcome the problem of training KANs slowly, we modify the implementation of original KANs to accelerate the training. Experiments on four real-world datasets show that the proposed KA2NCD exhibits better performance than traditional CDMs, and the proposed KA2NCD still has a bit of performance leading even over the existing neural CDMs. More importantly, the learned structures of KANs enable the proposed KA2NCD to hold as good interpretability as traditional CDMs, which is superior to existing neural CDMs. Besides, the training cost of the proposed KA2NCD is competitive to existing models.

Endowing Interpretability for Neural Cognitive Diagnosis by Efficient Kolmogorov-Arnold Networks

TL;DR

The paper tackles the interpretability gap in neural cognitive diagnosis models by introducing KA2NCD, which uses efficient Kolmogorov-Arnold networks to replace opaque MLP components or form a two-level KAN-based aggregation. It develops two deployment modes: KA2NCD-native that substitutes FC layers with KANs, and KA2NCD-e/KA2NCD-kan that processes embeddings with a two-level KAN architecture, combining lower-level and upper-level KANs to produce predictions. The authors implement training-time accelerations for KANs and validate KA2NCD across four real-world educational datasets, showing improved or competitive predictive performance and interpretability on par with traditional CDMs, with reasonable training costs. Overall, KA2NCD demonstrates that structured KAN-based architectures can provide transparent, interpretable cognitive diagnostics without sacrificing accuracy, enabling more trusted personalized education applications.

Abstract

In the realm of intelligent education, cognitive diagnosis plays a crucial role in subsequent recommendation tasks attributed to the revealed students' proficiency in knowledge concepts. Although neural network-based neural cognitive diagnosis models (CDMs) have exhibited significantly better performance than traditional models, neural cognitive diagnosis is criticized for the poor model interpretability due to the multi-layer perception (MLP) employed, even with the monotonicity assumption. Therefore, this paper proposes to empower the interpretability of neural cognitive diagnosis models through efficient kolmogorov-arnold networks (KANs), named KAN2CD, where KANs are designed to enhance interpretability in two manners. Specifically, in the first manner, KANs are directly used to replace the used MLPs in existing neural CDMs; while in the second manner, the student embedding, exercise embedding, and concept embedding are directly processed by several KANs, and then their outputs are further combined and learned in a unified KAN to get final predictions. To overcome the problem of training KANs slowly, we modify the implementation of original KANs to accelerate the training. Experiments on four real-world datasets show that the proposed KA2NCD exhibits better performance than traditional CDMs, and the proposed KA2NCD still has a bit of performance leading even over the existing neural CDMs. More importantly, the learned structures of KANs enable the proposed KA2NCD to hold as good interpretability as traditional CDMs, which is superior to existing neural CDMs. Besides, the training cost of the proposed KA2NCD is competitive to existing models.
Paper Structure (23 sections, 21 equations, 7 figures, 3 tables)

This paper contains 23 sections, 21 equations, 7 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminaries and Related Work
  3. Preliminaries of Cognitive Diagnosis Task
  4. Related Work on Cognitive Diagnosis
  5. Traditional CDMs
  6. Neural CDMs
  7. Kolmogorov-Arnold Networks (KANs)
  8. The Proposed KA2NCD
  9. Manner 1: KA2NCD-native
  10. Manner 2: KA2NCD-e and KA2NCD-kan
  11. The embedding module
  12. The main module
  13. Model Training and Implementation Modification
  14. Experiments
  15. Experimental Settings
  16. ...and 8 more sections

Figures (7)

  • Figure 1: Illustration of cognitive diagnosis process of two students.
  • Figure 2: The Overview of the proposed KA2NCD, containing two implementation manners.
  • Figure 3: Visualization of NCD+ (upper)and KaNCD+ (lower) on JunYi datasets.
  • Figure 4: Overview-Visualization of KANs in NCD+ (upper)and KaNCD+ (lower) on the SLP datasets.
  • Figure 5: Visualization of pruned KANs in NCD+ (upper)and KaNCD+ (lower) on the SLP datasets.
  • ...and 2 more figures