New dataset targets emotion detection in Vietnamese social media

Researchers have introduced ViTOED, a new dataset designed for target-oriented emotion detection specifically within Vietnamese social media. This dataset contains over 10,000 user comments with more than 21,000 manually annotated opinion quadruples, capturing Vietnamese-specific linguistic nuances like implicit sources and targets. A baseline model utilizing structured sentiment graphs and various Vietnamese pre-trained language models was evaluated, revealing challenges in span detection and relation extraction for these tasks. AI

IMPACT Enables deeper analysis of user emotions toward entities in Vietnamese social media, highlighting challenges for current language models.

RANK_REASON The item describes a new dataset and associated research paper for a specific NLP task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• arXiv
• Hugging Face
• social media
• structured sentiment graphs
• Target-Oriented Emotion Detection
• Vietnamese
• ViTOED

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →