Elena-Simona Apostol

Awakened at CheckThat! 2026: X-IBIS: an Iterative Bi-Encoder Fine-Tuning with Hard-Negative Mining for Cross-Lingual Scientific Source Retrieval

IROH: Insightful Ranking Of Humor using Multi-Stage Hybrid Retrieval with Rationale-Distilled LLM Judges for JOKER 2026 Track Task 1 English

Our team, VANGUARD, presents IROH (Insightful Ranking of Humor), a three-stage retrieval system for JOKER Task 1 English at CLEF 2026, achieving first place on the leaderboard with 0.6347 MAP. Our pipeline combines hybrid sparse-dense retrieval, cross-encoder reranking, and a LoRA-adapted Large Language Model judge ens...

Through the Eyes of the Beholder: Biometric and Demographic Conditioning for Multimodal Sexism Detection

The VANGUARD team proposes a human-centered multimodal framework that analyses and incorporates the psychological and demographic characteristics of human annotators into the detection pipeline and releases the full pipeline and analysis to support reproducible human-centered modeling.

Cross-lingual Relation Extraction with Large Language Models: Zero-Shot, Few-Shot, and Fine-Tuned Evaluation on Romanian

The results show that Romanian incurs a 3 to 5 percentage point drop relative to English in prompt-only settings, that few-shot prompting provides marginal gains over zero-shot, and that QLoRA fine-tuning improves macro F1-Score by more than 22 percentage points in both languages while reducing the cross-lingual gap fr...