Papers by Ahmed Haj Ahmed
Disentangling Linguistic Relatedness from Task Alignment in Cross-Lingual Transfer (2026.acl-srw)
Copied to clipboard
| Challenge: | Large language models (LLMs) have advanced natural language processing, yet their benefits remain concentrated in English and a small number of high-resource languages. |
| Approach: | They fine-tuned large language models (4B–671B parameters) on Arabic and evaluated zero-shot reading comprehension on Semitic languages and non-Semitic controls. |
| Outcome: | The results show that models with weak baselines improve across all languages, whereas strong-baseline models show only marginal gains regardless of language family. |
AUDITA: A New Dataset to Audit Humans vs. AI Skill at Audio QA (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing audio question answering benchmarks emphasize sound event classification or caption-grounded queries. |
| Approach: | They propose a large-scale, real-world audio question answering benchmark to evaluate audio reasoning beyond surface-level acoustic recognition. |
| Outcome: | The proposed model achieves 32.13% accuracy while demonstrating comprehension of audio . state-of-the-art models perform poorly, with average accuracy below 8.86%. |
CULEMO: Cultural Lenses on Emotion - Benchmarking LLMs for Cross-Cultural Emotion Understanding (2025.acl-long)
Copied to clipboard
Tadesse Destaw Belay, Ahmed Haj Ahmed, Alvin C Grissom Ii, Iqra Ameer, Grigori Sidorov, Olga Kolesnikova, Seid Muhie Yimam
| Challenge: | Existing emotion benchmarks rely on keyword-based emotion recognition, overlooking cultural dimensions required for emotion understanding. |
| Approach: | They propose a benchmark to evaluate culturally-aware emotion prediction across six languages. |
| Outcome: | The proposed benchmark evaluates state-of-the-art LLMs on culture-aware emotion prediction and sentiment analysis tasks. |