NTCIR

📅 Registration deadline: June 1, 2026 (AoE) 🔗 Task: modelretrieval-1.github.io (site has moved!) 📝 Register: research.nii.ac.jp/ntcir/ntcir-19… #NTCIR19 #ModelSelection #MLOps (4/4)

You may (or may not) actually fine-tune models — any approach that predicts rankings without exhaustive experimentation is valid. But your methodology must be documented in your paper. Ground truth: organizer-run fine-tuning results. (3/4)

🧵 Advances in AI have led to a proliferation of pre-trained models across many domains. But selecting the right model for a new task remains costly and time-consuming. #NTCIR19 ModelRetrieval frames this challenge as an information retrieval problem. (1/4)

📅 Registration deadline: July 10, 2026 (AoE) 🔗 Task: sites.google.com/nlg.csie.ntu.e… 📧 finarg@nlg.csie.ntu.edu.tw 📝 Register: research.nii.ac.jp/ntcir/ntcir-19… #NTCIR19 #FinNLP #ArgumentMining (5/5)

Organized by AIST Japan, Academia Sinica Taiwan, and the University of Passau Germany. (4/5)

🧵 "We expect revenue to grow 20% next quarter." Is that a well-argued forecast? #NTCIR19 FinArg-3 asks NLP systems to assess the quality of financial forward-looking statements — across three domains. (1/5)

📅 Registration deadline: June 1, 2026 (AoE) 🔗 Task: sites.google.com/view/dagri/ 📦 Sample data: github.com/ntcir-dagri/DA… 📝 Register: research.nii.ac.jp/ntcir/ntcir-19… #NTCIR19 #DocumentAI #AgriculturalTech (4/4)

The preliminary test dataset is already on GitHub: github.com/ntcir-dagri/DA…. Final dataset drops in July 2026. Two orientation meetings were held (Feb & Apr). Start exploring now. (3/4)

🧵 Agricultural expertise is locked in complex PDFs — tables with implicit structure, figures with domain jargon. #NTCIR19 DAGRI wants AI to unlock it. Three subtasks, and a dataset already available. (1/4)

📅 Registration deadline: June 1, 2026 (AoE) 🔗 Task: sites.google.com/view/ntcir19re… 💬 Community: groups.google.com/g/regcom 📝 Register: research.nii.ac.jp/ntcir/ntcir-19… #NTCIR19 #ESG #RegTech (4/4)

The challenge: ESG reports are semi-structured, mix text and visual elements, and vary greatly across regions and industries. The benchmark spans 6 languages and 6 countries. The organizing team has prior work at SemEval (PromiseEval-2025) and FinNLP. (3/4)

🧵 ESG reporting is now a regulatory requirement in many countries. Manually checking compliance across 6 languages and 6 countries is impossible at scale. #NTCIR19 RegCom is building the automation. (1/4)

📅 Registration deadline: July 10, 2026 (AoE) 🔗 Task: huggingface.co/spaces/THUIR/A… 📝 Register: research.nii.ac.jp/ntcir/ntcir-19… #NTCIR19 #LLMEval #NLProc (5/5)

This is the 2nd edition following AEOLLM at NTCIR-18. The leaderboard and dataset are hosted on HuggingFace. Organized by the THUIR lab at Tsinghua University. (4/5)

🧵 LLMs are now writing long-form "deep research" reports. Evaluating them at scale is a major unsolved problem. #NTCIR19 AEOLLM-2 wants your system to solve it. (1/5)