How AI Helps GOAT Keep Sneakerheads a Step Ahead - Ep. 79

AI Techniques

GOAT Group employs advanced AI techniques to tackle the complex issue of sneaker authentication. Michael Hall explains that the company has amassed a vast dataset of authentic sneakers, including detailed photos and measurements, to aid in this process. However, the challenge lies in the scarcity of counterfeit samples, prompting a shift towards unsupervised learning and human-in-the-loop machine learning.

We really had to start going to more of an outlier detection, unsupervised learning approach, and we've really utilized human in the loop machine learning.

--- Michael Hall

This approach helps prioritize manual verification efforts on sneakers suspected to be fake, enhancing the overall efficiency of the authentication process 1 2.

   

Human Expertise

Manual authentication remains a crucial component in verifying sneaker authenticity at GOAT. Michael Hall highlights the expertise of manual authenticators who can identify intricate details, such as stitching patterns and material quality, that are often indistinguishable to the untrained eye. Despite initial attempts to automate this process, the complexity and precision required have reinforced the need for human expertise.

We have manual authenticators who get the shoe in their hands. And these are domain experts.

--- Michael Hall

This human element ensures that even the most convincing counterfeits are detected, maintaining the integrity of the marketplace 3 2.