SemEval-2027 Task 9 opens: multimodal news framing analysis across 4 languages
preslav_nakov · x · 2026-10-16
SemEval-2027 Task 9 "Analysis of Multimodal Framing in the News" is now open for registration, studying how news text and images systematically reinforce or contradict each other in shaping reader interpretation.
- Two subtasks: paragraph-level fine-grained text framing classification, and multimodal image framing evaluation under oracle and blind contexts.
- Four languages: English, Bulgarian, European Portuguese, Russian; teams may enter any subtask-language pair.
- Benchmark: built on the 14-class Card et al. media frame taxonomy.
- Evaluation window: January 10–31, 2027.
The task targets media profiling, cross-country comparison of journalistic practices, and detection of manipulation attempts, with an explicit text-analysis → image-analysis multimodal reasoning pipeline.
More from Research
- FBI method trains for Pass@K with one sample per update at 12.3% of the cost — EmtiyazKhan · 2026-10-16
- Berkeley researchers sample humanoid robot policies directly from weight space with diffusion — berkeley_ai · 2026-10-16
- OneWM: unified causal training shows world and action modeling improve each other — IAAI233 · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16
- 5.8M-param Speech Event Detector adds laughs, breaths and背景 sounds to Whisper transcripts — Ayush Kumar Bar · 2026-10-16
- Huawei's SUPRA-Tex turns one image into relightable 3D materials with WYSIWYG fidelity — Chen Yang · 2026-10-16