Presented and Exhibited at NVIDIA Financial AI Meet-up with Macnica
NVIDIA Financial AI Meet-up with Macnica
— Autonomous Evolution of Generative AI: Balancing "Data Sovereignty" and "Accuracy" —
On Thursday, April 17, 2026, the "NVIDIA Financial AI Meet-up with Macnica" was held at the Global Business Hub Tokyo (Grand Cube 3F, Otemachi Financial City), organized by Macnica, Inc., with the cooperation of Mitsubishi Estate Co., Ltd. and NVIDIA Japan.
This in-person seminar brought together approximately 100 professionals from financial institutions seeking to advance industry-specific, task-specific, and proprietary AI development. Leading experts shared insights on open-source model utilization, secure generative AI implementation for sensitive data, and the in-house development of AI systems.
Featured on the Official NVIDIA Blog
Following the event, NVIDIA Japan's official blog published an event report:
▶ The Future of Financial Services Shaped by AI: NVIDIA's Vision of an Intelligence-Centric Bank (Japanese article)
(NVIDIA Japan Blog, May 14, 2026)
The article highlights Shotaro Minami's presentation on "an LLM that mathematically guarantees information quality," as well as the lab's broader research framework — built on a theoretical foundation that distinguishes first-hand information from N-th hand information — and its use of NVIDIA tools.
Presentation by Our Lab
Shotaro Minami (Associate Professor, Kyoto University Graduate School of Management) delivered a presentation on the following topic:
"Information Quality Assurance LLM Transforming Finance: Four Operating Systems and the Theory Behind Them"
Research Exhibition Booth
In addition to the keynote, lab members staffed a research exhibition booth where they introduced the lab's research activities to attendees. Through active discussions with industry professionals focused on practical financial AI applications, new avenues for academia-industry collaboration were explored.
Agenda
| Time | Topic | Organization |
|---|---|---|
| 14:15–14:45 | Global Investment Trends in AI Factory and Open Source Integration | NVIDIA |
| 14:55–15:15 | Realizing a Finance-Specific Vertical LLM through Fine-Tuning | Rakuten Group, Inc. |
| 15:15–15:35 | Development of Enterprise Document-Specialized Language Models Using Nemotron and NeMo Data Designer | First Accounting Inc. |
| 15:45–16:05 | Efficient and High-Accuracy Construction of Industry/Task-Specific LLMs: A Case Study in Financial Regulation Compliance | Nomura Research Institute, Ltd. |
| 16:05–16:25 | Business-Led Generative AI Utilization in On-Premises Environments | Ricoh Company, Ltd. |
| 16:35–16:55 | Real-Environment Validation with On-Premises GPU × NVIDIA NIM | Daiwa Institute of Research Ltd. |
| 16:55–17:15 | Information Quality Assurance LLM Transforming Finance: Four Operating Systems and the Theory Behind Them | Kyoto University Graduate School of Management |
| 17:25–17:45 | Building a Finance-Specific LLM Using Open-Weight Models: Evaluation via Banking Practice Tests and Approaches to Improving Accuracy | Mizuho Financial Group, Inc. |
| 17:45–18:05 | Behind the Scenes of MUFG-KDDI Collaboration 2.0 | KDDI Corporation |
Event Overview
| Item | Details |
|---|---|
| Date & Time | April 17, 2026 (Thu) 14:00–20:00 (Reception from 13:30) |
| Venue | Global Business Hub Tokyo (Grand Cube 3F, Otemachi Financial City) |
| Organizer | Macnica, Inc. |
| Cooperation | Mitsubishi Estate Co., Ltd. / NVIDIA Japan |
| Capacity | Approx. 100 |