1 minute read

Finding the ‘Right’ Data to Train your AI Is Getting Easier

A recent article in The Register explored the increasing role of humans and how they come into play when training AI.

The article, written by Danny Bradbury, noted that there is “an expanding ecosystem of humans in the machine-learning feedback loop who keep the machines on track.”

Not surprisingly, Alegion was cited as a prime example of this trend. The company possesses a legion of skilled, on-demand workers from around the world, enabling it to consistently deliver quality and scale to its customers. 

This is important, suggested Bradbury, not only at the outset when the training data is created, but also because humans are pivotal when they provide feedback “back into the training set to help the computer refine its own model.”

He used Alegion as an example of a company that has perfected a formula, highlighting the algorithm it built for one of its customers to detect damage on car body panels.

Crowdsourced workers outline the damage on hundreds of thousands of images of car body panels, and then classify them with jobs sent to Amazon's Mechanical Turk, Bradbury writes.

"What they need are examples of graded pictures, so they have a clear classification taxonomy for mild, moderate and severe damage," Alegion’s Chief Technology Officer Chip Ray told Bradbury.

Alegion helps that company, and many others, meet such needs. With accurate data, these companies can move forward and more fully realize the powerful promise of AI.

Read the article from The Register here: https://www.theregister.co.uk/2018/04/23/training_data_for_ai/

Learn More About Our Annotation Solutions