BGM-HAN: A Hierarchical Attention Network for Accurate and Fair Decision Assessment on Semi-Structured Profiles

2025-07-23

AI Summary

High-risk human decision-making—such as university admissions—is susceptible to implicit cognitive biases, undermining fairness and long-term outcomes. To address this, we propose BGM-HAN, a novel model grounded in byte-pair encoding and gated multi-head hierarchical attention, specifically designed for joint modeling of semi-structured application materials (i.e., free-text documents augmented with structured fields). This is the first approach to simultaneously achieve high semantic fidelity and strong interpretability in this setting. Evaluated on real-world admissions data, BGM-HAN significantly outperforms conventional machine learning baselines and large language models in both predictive accuracy and group-level fairness metrics—including calibration and equal opportunity. The model’s architecture enables transparent attribution of decisions to specific textual and structural features, facilitating auditability and stakeholder trust. All code is publicly available.

Technology Category

• Machine Learning: Ethics, Bias, and Fairness
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Human decision-making in high-stakes domains often relies on expertise and heuristics, but is vulnerable to hard-to-detect cognitive biases that threaten fairness and long-term outcomes. This work presents a novel approach to enhancing complex decision-making workflows through the integration of hierarchical learning alongside various enhancements. Focusing on university admissions as a representative high-stakes domain, we propose BGM-HAN, an enhanced Byte-Pair Encoded, Gated Multi-head Hierarchical Attention Network, designed to effectively model semi-structured applicant data. BGM-HAN captures multi-level representations that are crucial for nuanced assessment, improving both interpretability and predictive performance. Experimental results on real admissions data demonstrate that our proposed model significantly outperforms both state-of-the-art baselines from traditional machine learning to large language models, offering a promising framework for augmenting decision-making in domains where structure, context, and fairness matter. Source code is available at: https://github.com/junhua/bgm-han.

Problem

• Detect cognitive biases in high-stakes decision-making
• Improve fairness and accuracy in university admissions
• Model semi-structured data for nuanced applicant assessment

Innovation

• Hierarchical Attention Network for decision assessment
• Byte-Pair Encoded Gated Multi-head Attention
• Multi-level representation for interpretability