LinguIUTics at PsyDefDetect: Iterative Imbalance-Aware Fine-tuning of Qwen3-8B for Psychological Defense Mechanism Classification
arXiv:2606.00647v1 Announce Type: new Abstract: Detecting psychological defense mechanisms in conversational text remains a challenging clinical NLP problem. For the PsyDefDetect 2026 shared task (nine-class utterance classification evaluated via macro F1), our team LinguIUTics achieves a macro F1-score of 0.3917 on the official positive-class leaderboard, ranking 4th out of 21 registered teams and improving over the Ministral-8B task baseline (31.48 macro F1) by 7.7 absolute points (24.4 percent relative). BERT-family encoders and zero-shot LLMs proved ineffective on rare classes due to sever
The continuous improvement in NLP models and the increasing focus on complex psychological AI applications are driving research in this specialized area.
This development indicates progress in AI's ability to interpret nuanced human language for clinical applications, potentially enhancing mental health diagnostics and support systems.
AI models are becoming more adept at identifying subtle psychological defense mechanisms in text, moving beyond general sentiment analysis to more clinically relevant interpretations.
- · Clinical NLP researchers
- · Mental health tech startups
- · AI developers specializing in healthcare
- · Traditional diagnostic methods without AI augmentation
- · General-purpose LLMs without specialized fine-tuning
Improved accuracy in detecting psychological states via conversational AI.
Development of more effective AI-driven support tools for mental health assessment and intervention.
Potential for AI to augment or even partially automate initial psychological screening processes, raising ethical and regulatory considerations.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL