00:10 — This episode is brought to you by the Cloud Wars Expo. This in-person event will be held June 28th to 30th at the Moscone Center in San Francisco, California.
00:35 — Synthetic data is annotated information that computer simulations or algorithms generate as an alternative to real-world data. It’s created in digital worlds rather than collected or measured in the real world.
01:20 — What are the practical use cases of synthetic data? For example, in Nigeria, data scientists have recognized that data used to train computer vision AI models were often based on western clothing styles. Because there weren’t any for African clothing styles, they used AI to generate a new data set from artificial images of African fashion.
01:54 — Another synthetic data use case is driverless cars and driving on virtual streets. This allows for AI training across different scenarios without the cost or risk of real streets and cars.
02:18 — Accelerating synthetic data adoption, MIT’s data AI lab provides open-source tools for creating different types of data to help people get started with synthetic data.
02:33 — The real purpose of synthetic data is to fill the gaps of real-world data, bias data, or data that’s sensitive in nature. Rather than waiting for events to happen in the real world, users can turn to these simulations and digital scenarios to create syntehtic data.
03:07 — While gaps exist in data and AI models, it still requires people to identify the issues to create a foundation of AI and apply synthetic data to fill the gaps.
Looking for real-world insights into AI? Subscribe to the Enterprise AI Impact channel: