Elith and NVIDIA release 9 million synthetic Japanese financial personas
Reported by Hugging Face and Unite.AI2 sources ↓
Elith and NVIDIA released a synthetic dataset of 9 million Japanese adult personas with demographic and household-finance attributes. It is available under CC BY 4.0 for commercial and non-commercial use.
Key points
- The dataset adds household finances and financial-literacy attributes to Japanese demographic profiles
- It contains 50 fields per record and covers all 47 prefectures
- The personas are synthetic, not records of real individuals
- A companion release contains 100,000 synthetic Japanese business documents labeled for personal information
Sources
Details
The dataset extends Nemotron-Personas-Japan, which had six million personas across 22 fields. The new release has nine million personas across 50 fields, covering adults aged 18 and older and more than 1,600 occupation categories. Elith and NVIDIA say it is the first open persona dataset at million-record scale to combine Japan’s official statistics with household-finance and financial-literacy data.
The release is licensed under CC BY 4.0. Intended uses include building and testing customer-service systems, simulating customer interactions, evaluating sales conversations and red-teaming financial AI. The dataset was developed using NVIDIA’s NeMo Data Designer and NeMo SDG-PGMs; the release announcement says financial institutions and an industry association provided design feedback.
The dataset draws on public statistics and survey data, including the Statistics Bureau’s 2019 National Survey of Family Income, Consumption and Wealth. A companion dataset, Nemotron-PII-Japan, was released the same day; it contains 100,000 synthetic Japanese business documents with labeled personal information for training and testing detection and redaction systems.
Written from coverage as of
Images
- Hugging Face
- Unite.AI
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh