VISHC at PsyDefDetect: Mitigating Data Scarcity in Psychological Defense Classification with Context-Aware Synthetic Augmentation

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Science & Research — Health & Medical Research, Social Sciences & Behavioral Studies · Depth: Advanced, quick

Summary

VISHC's work for the PsyDefDetect shared task (BioNLP@ACL 2026) addresses the challenge of classifying psychological defense mechanisms (PDMs) from text, a task hindered by severe data scarcity and class imbalance. The team proposes a context-aware synthetic augmentation framework integrated with a hybrid classification model. This model combines contextual language representations with basic clinical features and utilizes 150 annotated defense items. Experiments revealed that the quality of definitions used in prompting directly impacts generation fidelity and subsequent classification performance. The VISHC method achieved an accuracy of 58.26%, marking a 40.25% improvement, and a macro-F1 of 24.62%, an increase of 15.99%, surpassing the DMRS Co-Pilot baseline. This establishes a robust foundation for psychologically grounded defense mechanism classification in settings with limited data.

Key takeaway

For NLP Engineers developing classification models in low-resource clinical domains, you should prioritize context-aware synthetic data augmentation. Ensure your generative prompts use high-quality, psychologically grounded definitions; this directly impacts model performance. Integrating basic clinical features with contextual language representations can significantly improve accuracy and F1 scores. This establishes a stronger baseline for challenging tasks like psychological defense mechanism classification.

Key insights

Context-aware synthetic augmentation, driven by definition quality, significantly improves psychological defense mechanism classification in data-scarce environments.

Principles

• Data scarcity severely hinders PDM classification.
• Generative augmentation requires psychological grounding.
• Prompt definition quality governs generation fidelity.

Method

A context-aware synthetic augmentation framework is combined with a hybrid classification model. This model integrates contextual language representations with basic clinical features and 150 annotated defense items.

In practice

• Employ context-aware synthetic augmentation.
• Integrate clinical features with language models.
• Ensure high-quality definitions for prompts.

Topics

• Psychological Defense Mechanisms
• Text Classification
• Data Augmentation
• Low-Resource NLP
• Clinical NLP
• Hybrid Models
• Prompt Engineering

Code references

• htdgv/CASA-PDC

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• Mitigating Data Scarcity in Psychological Defense Classification with Context-Aware Synthetic Augmentation — Context-aware synthetic data augmentation with theory-driven prompts significantly improves psychological defense mechanism classification in low-resource settings.
• Mitigating Data Scarcity in Psychological Defense Classification with Context-Aware Synthetic Augmentation — Context-aware synthetic augmentation significantly improves psychological defense mechanism classification in low-resource settings.
• Low-resource Language Discrimination Towards Chinese Dialects with Transfer learning and Data Augmentation — Transfer learning and data augmentation effectively overcome scarce annotation resources for low-resource Chinese dialect discrimination.
• LinguIUTics at PsyDefDetect: Iterative Imbalance-Aware Fine-tuning of Qwen3-8B for Psychological Defense Mechanism Classification — Iterative imbalance-aware fine-tuning of LLMs, combined with augmentation and post-processing, significantly improves rare-class NLP performance.
• IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages — AudioLLMs' actual context utilization from prompts is ambiguous, necessitating explicit evaluation benchmarks like IndicContextEval to reveal true grounding.
• 5 Examples of the Importance of Context-Sensitivity in Data-Centric AI — Contextual data is crucial for accurate NLP models, preventing misinterpretations and improving performance.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.