better data, a Singapore-based startup that uses programmable synthetic data to keep real data safe, announced today that it has raised $1.55 million. The oversubscribed seed round was led by Investible with participation from Franklin Templeton, Xcel Next, Singapore University of Technology and Design, Bon Auxilium, Tenity, Plug and Play and Entrepreneur First.
With the goal of making data sharing faster and more secure as data protection regulations tighten around the world, the startup is led by CEO and Chief Technologist Kevin Yee, Ph.D., Uzair Javaid. Founded in 2021 by Dr. The company currently has R&D partnerships with his two major universities in Singapore and the US (who they are unable to disclose), and customers include Shanghai Pudong Development Bank.
Betterdata says it relies on generative AI and privacy engineering instead of traditional data sharing methods that use data anonymization to destroy data.
Yee told TechCrunch that programmatically synthesized data uses generative models such as the generative adversarial model used in deepfakes, the transformer used in ChatGPT, and deep learning models such as the diffusion model used in stable diffusion. and explained that we are creating and extending a new dataset.
These synthetic datasets have similar properties and structure to real-world data without revealing sensitive or private information about individuals.
“The idea is to create fictional versions of real datasets that can be safely used for a variety of purposes, including protecting sensitive data, mitigating bias, and improving machine learning models,” he said.
Programmatically synthesized data helps developers in many ways. Some examples include protecting sensitive data, complying with data protection regulations such as GDPR and HIPAA, increasing data availability across teams, creating data for training, testing, and validating machine learning models, underestimation for example, addressing data imbalance issues by increasing the number of recorded groups. Or class.
Betterdata funding will be used for product launches and to power our programmable synthetic data technology stack, including support for single-table, multi-table, and time-series datasets. Yee explains that these are different variations of tabular datasets, the main difference being their structure and the problems they were created to address.
For example, single-table datasets focus on stand-alone tables, multi-table datasets aim to consider the relationships between multiple tables, and time series datasets Handle collected data.
Betterdata also plans to hire more people, including sales and marketing employees, and expand beyond Singapore into the Asia-Pacific region in the next year or two.
In Investible’s investment statement, Principal Kyle Regal said: Through its powerful platform, Betterdata generates synthetic data that mimics real-world data without compromising quality and privacy, helping businesses meet global compliance and privacy laws at scale. ”