Betterdata, a Singaporean start-up that uses programmable synthetic data to secure real data announced today that they had raised $1,55million. The seed round led Investible received a high level of interest. Franklin Templeton and Xcel Next, as well as Singapore University of Technology and Design, were also involved. Bon Auxilium as well as Tenity participated. Plug and Play and Entrepreneur First also participated.
Dr. Uzairjavaid, the founder, and Kevin Yee the chief technologist founded the startup in 2021 with the goal of making data sharing faster and safer as data protection laws around the world increase. The company has established research and development partnerships with two major US and Singaporean universities. These universities cannot be publicly revealed.
Betterdata claims it is different from traditional methods of data sharing, which use data anonymization and privacy engineering for destruction.
Yee explained TechCrunch programmatic synthesized datasets use generative modeling, such as deep learning models and generative adversarial modelling used in deepfakes. ChatGPT uses transformers. In stable diffusion, diffusion models are used.
These synthetic data sets have the same structure and characteristics as real information, without revealing any sensitive or personal data about individuals.
He said that he was creating a fictionalized dataset which could be utilized for various purposes, such as safeguarding sensitive data and reducing bias.
The synthetic data generated programmatically is useful for developers in many ways. Some examples include helping to protect sensitive data and complying with data privacy regulations such as GDPR or HIPAA. They can also increase the data availability between teams, create more data to train, test and validate machine learning models, as well as address data imbalances by creating more records or class for underrepresented groups.
Betterdata will use this funding to launch their product and enhance their programmable synthesized tech stack. This includes support for multi-table datasets and time-series. Yee explains what the differences are between these tabular data sets, mainly the structure and the problem they are designed to address.
Multi-table datasets are designed to examine relationships between multiple tables. Time-series datasets are those that deal with data collected at a specific time.
Betterdata also plans to hire additional staff, including marketing, sales, and expand beyond Singapore into other parts of Asia-Pacific Region in the next one-to-two year.
Khairu Rjal is the principal at Investible. He said, “Betterdata solves one of the biggest challenges facing the AI industry: a lack high-quality data that meets privacy standards. Through its powerful platform, which mimics real-world data, Betterdata generates synthetic information, without compromising quality or privacy. This helps companies comply with privacy and compliance laws around the world.