Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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.

Srikar Kashyap Pulipaka• 2026

Related benchmarks

TaskDatasetResultRank
Polarization detectionSemEval Task 9 2026 (test)
Macro-F181.2
17
Subtask 1POLAR amh (test)
Score0.8
2
Subtask 1POLAR hin (test)
Score0.824
2
Subtask 1POLAR swa (test)
Score0.811
2
Subtask 1POLAR arb (test)
Score84.8
2
Subtask 1POLAR ben (test)
Score83.7
2
Subtask 1POLAR deu (test)
Score72.8
2
Subtask 1POLAR eng (test)
Score81.8
2
Subtask 1POLAR fas (test)
Score82.8
2
Subtask 1POLAR hau (test)
Score80
2
Showing 10 of 24 rows

Other info

Follow for update