Explicit Knowledge-Guided In-Context Learning for Early Detection of Alzheimer's Disease

2025-11-09

AI Summary

Early detection of Alzheimer’s disease (AD) from narrative clinical transcripts faces challenges including poor out-of-distribution (OOD) generalization, unstable inference under low-resource conditions, and task mismatch. To address these, we propose EK-ICL—a novel in-context learning (ICL) framework that explicitly integrates domain knowledge for the first time: confidence-scoring guides exemplar selection; feature-based parsing enhances task identification; and label-word replacement improves semantic alignment. Furthermore, EK-ICL fuses outputs from small language models, syntactic parsing features, dependency-structure-aware retrieval, and ensemble prediction to significantly boost ICL robustness on clinical text. Evaluated across three AD datasets, EK-ICL consistently outperforms state-of-the-art fine-tuning and ICL baselines. Our results demonstrate that explicit knowledge modeling is critical for reliable, low-resource clinical reasoning—establishing a new paradigm for knowledge-enhanced ICL in healthcare NLP.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Knowledge Acquisition from the Web

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Detecting Alzheimer's Disease (AD) from narrative transcripts remains a challenging task for large language models (LLMs), particularly under out-of-distribution (OOD) and data-scarce conditions. While in-context learning (ICL) provides a parameter-efficient alternative to fine-tuning, existing ICL approaches often suffer from task recognition failure, suboptimal demonstration selection, and misalignment between label words and task objectives, issues that are amplified in clinical domains like AD detection. We propose Explicit Knowledge In-Context Learners (EK-ICL), a novel framework that integrates structured explicit knowledge to enhance reasoning stability and task alignment in ICL. EK-ICL incorporates three knowledge components: confidence scores derived from small language models (SLMs) to ground predictions in task-relevant patterns, parsing feature scores to capture structural differences and improve demo selection, and label word replacement to resolve semantic misalignment with LLM priors. In addition, EK-ICL employs a parsing-based retrieval strategy and ensemble prediction to mitigate the effects of semantic homogeneity in AD transcripts. Extensive experiments across three AD datasets demonstrate that EK-ICL significantly outperforms state-of-the-art fine-tuning and ICL baselines. Further analysis reveals that ICL performance in AD detection is highly sensitive to the alignment of label semantics and task-specific context, underscoring the importance of explicit knowledge in clinical reasoning under low-resource conditions.

Problem

• Detecting Alzheimer's Disease from narrative transcripts using LLMs under OOD conditions
• Addressing task recognition failure and suboptimal demonstration selection in ICL
• Resolving semantic misalignment between label words and clinical task objectives

Innovation

• Integrates explicit knowledge for stable reasoning
• Uses confidence scores and parsing features
• Implements label replacement and ensemble prediction