Gladiators at #SMM4H–HeaRD 2026: Multi-Seed XLM-RoBERTa Ensemble with Focal Loss and Per-Language Threshold Optimization for Multilingual Adverse Drug Event Detection
Year: 2026
Type: conference-paper
Abstract: This paper describes the Gladiators system for Task 1 of the SMM4H 2026 shared task on binary classification of adverse drug event (ADE) mentions in multilingual social media posts.Our system fine-tunes three XLM-RoBERTalarge models with different random seeds using focal loss (α=0.75, γ=2.0) and 3× positive oversampling, then averages their predicted probabilities and applies per-language threshold optimization.On the development set, our ensemble achieves a pooled binary F1 of 0.7505.On the official test set-which introduced surprise Farsi comprising 35.5% of samples-our system achieves F1 = 0.6039, above the competition mean (0.5465) and median (0.5798).We evaluated eleven approaches, document key negative results including catastrophic translation augmentation failures, and analyze precision-recall trade-offs under severe class imbalance.Post-evaluation, a sixmodel cross-regime ensemble improved dev F1 to 0.7585.
Author: Ankit Kumar Singh
Language: English
Subfield: Toxicology
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W7166878558