Tech
AI startup Datagen uses computer vision to reduce bias in algorithms. Here’s the 18-slide pitch deck it used to raise $50 million.
Datagen
subscribers.
Become an Insider
and start reading now.
Have an account? .
- AI data startup Datagen recently raised a $50 million Series B round.
- The firm gives customers access to synthetic data on which to train artificial intelligence algorithms.
- You can see Datagen’s fundraising pitch deck below.
Datagen, an Israeli computer vision startup, has raised $50 million in a Series B funding round backed by Scale Venture Partners and Spider Capital.
Computer vision is a growing sub field of AI, focused on training machines to extract huge amounts of information from images and videos.
Datagen offers up synthetic data from the training of AI models. All artificial intelligence tech is trained on information — but where that information comes from and what it comprises has implications for accuracy and bias. Synthetic data, simulated by a computer rather than from the real world, offers a possible solution.
subscribers.
Become an Insider
and start reading now.
Have an account? .
In a recent study, reviewed on OpenReview, MIT researchers used machine learning to generate synthetic data to train another model for vision-related tasks. They found a model trained on synthetic data matched or outperformed a model trained on real-world data.
Datagen says customers who rely on its data generation use it to inform everything from robotics to virtual reality and self-driving vehicles.
The startup wants to be the one-stop shop for companies’ in-house computer vision teams. Andy Vitus, partner at Datagen backer Scale Venture Partners, said providing clean data for AI training could be “the most lucrative piece of the machine learning puzzle.”
Alongside Scale Venture Partners and Spider Capital, Datagen’s latest funding round saw renewed investment from existing backers TLV Partners and Viola Ventures.
Check out Datagen’s pitch deck below:
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen
Datagen