PSK at SemEval-2026 Task 9: Multilingual Polarization Detection Using Ensemble Gemma Models with Synthetic Data Augmentation

• URL: http://arxiv.org/abs/2605.05159v1
• Date: Wed, 06 May 2026 17:29:14 GMT
• Title: PSK at SemEval-2026 Task 9: Multilingual Polarization Detection Using Ensemble Gemma Models with Synthetic Data Augmentation
• Abstract summary: We present our system for SemEval-2026 Task 9: Multilingual Polarization Detection, a binary classification task spanning 22 languages.<n>Our approach fine-tunes separate Gemma3 models per language using Low-Rank Adaptation (LoRA), augmented with synthetic data generated by a large language model (LLM)<n>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.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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 papers

• HABIB_TAZ at SemEval-2026 Task 11: Disentangling Formal Logic from Content via Synthetic Training and Multi-Objective Optimization [0.0]
  We present our system for SemEval-2026 Task 11, which evaluates the ability of models to disentangle formal logic from content across 12 languages.<n>We address this challenge using mDeBERTa-v3 networks fine-tuned on a synthetic, rule-based dataset of syllogistic schemes.<n>Our training pipeline employs a multi-objective loss function combining Adaptive Group Distributionally Robust Optimization (DRO), a scheduled differentiable bias penalty, and KL-Divergence consistency regularization.
  arXiv  Detail & Related papers  (2026-07-15T20:24:26Z)
• MKJ at SemEval-2026 Task 9: A Comparative Study of Generalist, Specialist, and Ensemble Strategies for Multilingual Polarization [0.0]
  We present a systematic study of multilingual polarization detection across 22 languages for SemEval-2026 Task 9 (Subtask 1).<n>Rather than enforcing a single universal architecture, we adopt a language-adaptive framework that switches between multilingual generalists, language-specific specialists, and hybrid ensembles based on development performance.
  arXiv  Detail & Related papers  (2026-04-23T07:34:26Z)
• Toward Generalized Cross-Lingual Hateful Language Detection with Web-Scale Data and Ensemble LLM Annotations [1.3912467776690247]
  We study whether large-scale unlabelled web data and LLM-based synthetic annotations can improve multilingual hate speech detection.<n>We apply continued pre-training to BERT models by continuing masked language modelling on unlabelled OWS texts.<n>We show that this yields an average macro-F1 gain of approximately 3% over standard baselines.
  arXiv  Detail & Related papers  (2026-03-18T13:57:23Z)
• Binary Token-Level Classification with DeBERTa for All-Type MWE Identification: A Lightweight Approach with Linguistic Enhancement [1.8429656136522097]
  We present a comprehensive approach for multiword expression (MWE) identification that combines binary token-level classification, linguistic feature integration, and data augmentation.<n>Our DeBERTa-v3-large model achieves 69.8% F1 on the CoAM dataset, surpassing the best results (Qwen-72B, 57.8% F1) on this dataset by 12 points while using 165x fewer parameters.
  arXiv  Detail & Related papers  (2026-01-27T08:42:54Z)
• Seed-X: Building Strong Multilingual Translation LLM with 7B Parameters [53.59868121093848]
  We introduce Seed-X, a family of open-source language models (LLMs) with 7B parameter size.<n>The base model is pre-trained on a diverse, high-quality dataset encompassing both monolingual and bilingual content across 28 languages.<n>The instruct model is then finetuned to translate by Chain-of-Thought (CoT) reasoning and further enhanced through reinforcement learning (RL) to achieve better generalization across diverse language pairs.
  arXiv  Detail & Related papers  (2025-07-18T03:19:43Z)
• DeMuX: Data-efficient Multilingual Learning [57.37123046817781]
  DEMUX is a framework that prescribes exact data-points to label from vast amounts of unlabelled multilingual data. Our end-to-end framework is language-agnostic, accounts for model representations, and supports multilingual target configurations.
  arXiv  Detail & Related papers  (2023-11-10T20:09:08Z)
• Breaking Language Barriers in Multilingual Mathematical Reasoning: Insights and Observations [59.056367787688146]
  This paper pioneers exploring and training powerful Multilingual Math Reasoning (xMR) LLMs. We construct the first multilingual math reasoning instruction dataset, MGSM8KInstruct, encompassing ten distinct languages. By utilizing translation, we construct the first multilingual math reasoning instruction dataset, MGSM8KInstruct, encompassing ten distinct languages.
  arXiv  Detail & Related papers  (2023-10-31T08:09:20Z)
• HFL at SemEval-2022 Task 8: A Linguistics-inspired Regression Model with Data Augmentation for Multilingual News Similarity [16.454545004093735]
  This paper describes our system designed for SemEval-2022 Task 8: Multilingual News Article Similarity. We proposed a linguistics-inspired model trained with a few task-specific strategies. Our system ranked 1st on the leaderboard while achieving a Pearson's Correlation Coefficient of 0.818 on the official evaluation set.
  arXiv  Detail & Related papers  (2022-04-11T03:08:37Z)
• Mixed-Lingual Pre-training for Cross-lingual Summarization [54.4823498438831]
  Cross-lingual Summarization aims at producing a summary in the target language for an article in the source language. We propose a solution based on mixed-lingual pre-training that leverages both cross-lingual tasks like translation and monolingual tasks like masked language models. Our model achieves an improvement of 2.82 (English to Chinese) and 1.15 (Chinese to English) ROUGE-1 scores over state-of-the-art results.
  arXiv  Detail & Related papers  (2020-10-18T00:21:53Z)
• Explicit Alignment Objectives for Multilingual Bidirectional Encoders [111.65322283420805]
  We present a new method for learning multilingual encoders, AMBER (Aligned Multilingual Bi-directional EncodeR) AMBER is trained on additional parallel data using two explicit alignment objectives that align the multilingual representations at different granularities. Experimental results show that AMBER obtains gains of up to 1.1 average F1 score on sequence tagging and up to 27.3 average accuracy on retrieval over the XLMR-large model.
  arXiv  Detail & Related papers  (2020-10-15T18:34:13Z)
• FiSSA at SemEval-2020 Task 9: Fine-tuned For Feelings [2.362412515574206]
  In this paper, we present our approach for sentiment classification on Spanish-English code-mixed social media data. We explore both monolingual and multilingual models with the standard fine-tuning method. Although two-step fine-tuning improves sentiment classification performance over the base model, the large multilingual XLM-RoBERTa model achieves best weighted F1-score.
  arXiv  Detail & Related papers  (2020-07-24T14:48:27Z)
• Learning to Learn Morphological Inflection for Resource-Poor Languages [105.11499402984482]
  We propose to cast the task of morphological inflection - mapping a lemma to an indicated inflected form - for resource-poor languages as a meta-learning problem. Treating each language as a separate task, we use data from high-resource source languages to learn a set of model parameters. Experiments with two model architectures on 29 target languages from 3 families show that our suggested approach outperforms all baselines.
  arXiv  Detail & Related papers  (2020-04-28T05:13:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.