Nürnberg NLP at PsyDefDetect: Multi-Axis Voter Ensembles for Psychological Defence Mechanism Classification
For researchers in computational psycholinguistics, this work provides a practical ensemble method for a low-agreement classification task, though it is an incremental application of existing techniques.
The authors tackled the problem of detecting psychological defence mechanisms in supportive conversations, where categories are ambiguous and raters have low agreement. Their multi-axis voter ensemble achieved an F1 of 0.420 on the hidden test set, ranking first among 21 teams.
Detecting levels of psychological defence mechanisms in supportive conversations is inherently ambiguous. In the PsyDefDetect shared task at BioNLP 2026 the eight positive defence categories share surface language and differ only in pragmatic function and trained raters reach only moderate inter-annotator agreement. On such a task the decisive lever is not a stronger single model but error independence, since any single representation will waver on the overlapping defence boundaries. We translate this insight into a 9-voter ensemble spanning three orthogonal axes: class granularity (all nine classes for the gatekeeper, only the eight defence classes for the specialists), training method (generative and discriminative) and base model. The system reaches $F1_{test}{=}.420$ on the hidden test set, placing first among 21 registered teams.