PSK at SemEval-2026 Task 9: Multilingual Polarization Detection Using Ensemble Gemma Models with Synthetic Data Augmentation
About
We present our system for SemEval-2026 Task 9: Multilingual Polarization Detection, a binary classification task spanning 22 languages. Our approach fine-tunes separate Gemma~3 models (12B and 27B parameters) per language using Low-Rank Adaptation (LoRA), augmented with synthetic data generated by a large language model (LLM). We employ three synthetic data strategies (direct generation, paraphrasing, and contrastive pair creation) using GPT-4o-mini, with a multi-stage quality filtering pipeline including embedding-based deduplication. We find that per-language threshold tuning on the development set yields 2 to 4\% F1 improvements without retraining. We also use weighted ensembles of 12B and 27B model predictions with per-language strategy selection. Our final system achieves a mean macro-F1 of 0.811 across all 22 languages, ranking 2nd overall of the participating teams, with 1st place finishes in 3 languages and top-3 in 8 languages. We also find that alternative architectures (XLM-RoBERTa, Qwen3) that showed strong development set performance suffered 30 to 50\% F1 drops on the test set, highlighting the importance of generalization.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Polarization detection | SemEval Task 9 2026 (test) | Macro-F181.2 | 17 | |
| Subtask 1 | POLAR amh (test) | Score0.8 | 2 | |
| Subtask 1 | POLAR hin (test) | Score0.824 | 2 | |
| Subtask 1 | POLAR swa (test) | Score0.811 | 2 | |
| Subtask 1 | POLAR arb (test) | Score84.8 | 2 | |
| Subtask 1 | POLAR ben (test) | Score83.7 | 2 | |
| Subtask 1 | POLAR deu (test) | Score72.8 | 2 | |
| Subtask 1 | POLAR eng (test) | Score81.8 | 2 | |
| Subtask 1 | POLAR fas (test) | Score82.8 | 2 | |
| Subtask 1 | POLAR hau (test) | Score80 | 2 |