Make Customers & Algorithms Happy: Bootstrapping Training Data with a Consumer Product
Deep learning needs a massive amount of training data. Collecting the relevant training data for a machine learning driven consumer product is a chicken egg problem. You need users to get real world data to train your algorithm and you need a working algorithm to provide a great experience to the users. In this talk we show how a consumer product can be used to generate the much need training data to solve the user problem.
Fashwell is a Swiss Fashion-Tech startup that's closing the gap between visual inspirational content in social media and commerce by making every image fashion shoppable."""
Matthias Dantone is the Co-founder and CEO of the fashion-tech company Fashwell AG. He received his PhD from the ETH in Zurich with leading results in the computer vision and machine learning community. Before starting his own company he worked for several small and big companies in the are of computer vision Google , Kooaba and Majachitra. Matthias is collecting sneakers and spent his free time on his bike or skies.