Researchers have developed a system to track and visualize election commitments made by political candidates, particularly in developing countries where accountability is often lacking. This work introduces the first gold-standard dataset for Nepali political texts and evaluates various AI models for extracting these commitments. LaBSE-LR and fastText-LSTM models showed strong performance, while the Qwen3_8B model demonstrated effectiveness in a few-shot setting. The system is integrated into a mobile application to enhance transparency and civic engagement. AI
IMPACT Enhances transparency and accountability in political processes through AI-driven analysis of campaign promises.
RANK_REASON Academic paper detailing a new dataset and model evaluation for election commitment mining. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- FastText + LSTM
- Gotit.pub
- Hugging Face
- LaBSE-LR
- Qwen3_8B
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →