It's been a while and the world has gotten hip with deep learning. We don't care much for fashion but we do care if people manage to apply them. Turns out we've found two such people who are kind enough to talk about their experiences. We've also found a great host [Optiver!] who is willing to let us use their new auditorium. All the more reason to meet up!
Warning: If you are already attending this meet up through the Amsterdam Machine Learning Group Meetup, no need to register here!
Talk 1 - Personalised recommendations by transfer learning of images - Pierre Gutierrez
For travel e-business companies, recommender systems are paramount. There is an increasing need to take into account all the user information to tailor the best product proposition. One of them is the content that the user actually sees: the visual of the product. When it comes to hostels, some people can be more attracted by pictures of the room, the building or even the nearby beach. In this talk, we will describe how we improved an e-business vacation retailer recommender system using the content of images. We’ll explain how to leverage open dataset and pre-trained deep learning models to derive user taste information. This transfer learning approach enables companies to use state of the art machine learning methods without having deep learning expertise.
Pierre Gutierrez is lead data scientist at Dataiku Labs in Paris, France. In the past few years he has been working on state of the arts Data Science and Machine learning problems in a large variety of sectors such as e-business, retail, insurance or telcos. He has experience in topics such as fraud detection, bot detection, recommender systems, or churn prediction. Pierre has a pragmatic approach to data science and strongly believes in the power of transfer learning in image, text and artificial intelligence.
Talk 2 - An Architecture for Intelligent Reporting at Scale - Stephen Helms
In this talk, I will discuss the architectural decisions and implementations we have used to build our automated reporting architecture. This includes the use of online (recursive) implementations of Bayesian statistics for estimating metrics and detecting trends and outliers, NoSQL databases for scalable storage and evolving data schemas, parallel execution across a Mesos cluster, and publication of the reports through REST APIs. All implemented in Python.
Stephen Helms started as a research scientist studying information processing in biological systems. Stephen now tries to make sense of equally complicated financial networks. He works on the Data team at Optiver, a market maker trading on exchanges around the world. Stephen is passionate about developing statistical machine learning algorithms and data visualizations to uncover hidden patterns and trends in real world data.