AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Identifying Causal Features in Graph Machine Learning
PayPal is using GCN's to learn embeddings for several kinds of use cases like collusion detection. The problem with graph, which I was alluding to earlier is the scale of the graph that we have at PayPal. We have millions of customers, billions of transactions happening every year. So to model all of that, to keep that graph up to date and train models on this kind of evolving graph can be a very challenging problem. There are so many great papers and researches coming out in this area that it's a continuous area of exploration for us.