Path-Specific Causal Reasoning for Fairness-aware Cognitive Diagnosis

TL;DR

This work tackles fairness issues in cognitive diagnosis by addressing how sensitive student attributes can bias predictions. It introduces PSCRF, a framework that uses path-specific causal reasoning to remove unfair pathways from sensitive attributes while retaining diagnosis-relevant signals, through a decoupled predictor, counterfactual reasoning, and a multi-factor fairness constraint. Empirical results on PISA-derived Australia and Brazil data show PSCRF improves fairness metrics (e.g., EO, $D_{disadv}^{under}$) without sacrificing diagnostic accuracy across multiple CD models, with ablation and parameter analyses confirming the contribution of each component. The approach offers a practical, theoretically grounded method to deploy fairer cognitive diagnosis in intelligent education and is extensible to other CD architectures.

Abstract

Cognitive Diagnosis~(CD), which leverages students and exercise data to predict students' proficiency levels on different knowledge concepts, is one of fundamental components in Intelligent Education. Due to the scarcity of student-exercise interaction data, most existing methods focus on making the best use of available data, such as exercise content and student information~(e.g., educational context). Despite the great progress, the abuse of student sensitive information has not been paid enough attention. Due to the important position of CD in Intelligent Education, employing sensitive information when making diagnosis predictions will cause serious social issues. Moreover, data-driven neural networks are easily misled by the shortcut between input data and output prediction, exacerbating this problem. Therefore, it is crucial to eliminate the negative impact of sensitive information in CD models. In response, we argue that sensitive attributes of students can also provide useful information, and only the shortcuts directly related to the sensitive information should be eliminated from the diagnosis process. Thus, we employ causal reasoning and design a novel Path-Specific Causal Reasoning Framework (PSCRF) to achieve this goal. Specifically, we first leverage an encoder to extract features and generate embeddings for general information and sensitive information of students. Then, we design a novel attribute-oriented predictor to decouple the sensitive attributes, in which fairness-related sensitive features will be eliminated and other useful information will be retained. Finally, we designed a multi-factor constraint to ensure the performance of fairness and diagnosis performance simultaneously. Extensive experiments over real-world datasets (e.g., PISA dataset) demonstrate the effectiveness of our proposed PSCRF.

Figures (6)

• Figure 1: A tiny example of general cognitive diagnosis.
• Figure 2: The general causal graph and commonly used counterfactual notations.
• Figure 3: The causal graph used in our proposed PSCRF.
• Figure 4: The overall framework of our proposed PSCRF
• Figure 5: PSCRF based on IRT with varying $w$