Jekaterina is a Research Engineer at Zalando, focusing on applying machine learning for payment fraud prediction. Jekaterina obtained a masters degree in bioinformatics from FU Berlin and worked in various research institutions across Europe such as the Charité Hospital in Berlin, the Centre for Genomics Regulations in Barcelona and at Manchester University.
The cold start problem is one of the biggest challenges when applying machine learning in industry. The goal is to build an accurate production-ready machine learning model when there is either no data or not enough data available to train a model - the so-called cold start. In this case methods like transfer learning prove to be valuable by using a model that was trained on a different but related task.
In her talk Jekaterina will guide through a real-world cold start problem she was working on in payment fraud detection at Zalando. The business case aims at building a payment fraud detection model for a new market, where there is no historical data available. The challenge is how to combine the data from existing markets in an elaborate way so that it generalises to a new market. Jekaterina will discuss several possible solutions such as (online) transfer learning and domain generalisation. Additionally, she will explore how some of the techniques can be extended to be used on a stream of unlabeled or limited labeled data.