XGBoost (https://xgboost.ai/) (https://github.com/dmlc/xgboost), one of the most popular Gradient Boost Tree Systems, is widely adopted as a solution for data science competitions, like Kaggle, and many production scenarios at companies like Uber, Airbnb, Netflix, etc.
Join us for an evening of tech talks and networking with speakers from Uber, Amazon, and Ant Financial around leveraging this powerful ML tool in production environments.
Topics & Speakers:
Machine Learning with Large-scale Telematics Data: Wayne Zhang and Gorkem Ozkaya [Uber]
The increasing availability of mobile phones with embedded GPS devices and sensors has spurred the use of vehicle telematics in recent years. Telematics provides information on a vehicle, such as the location, speed, and movement. Additional insights can be learned from combining vehicle telematics with other spatial data. In this talk, we will explain how it is possible to transform massive and noisy telematics datasets into a structured form using a combination of manual feature engineering and deep learning approaches and then leveraging XGBoost for a set of classification and regression problems.
Learning Deep Forest Using XGBoost: Yin Lou [Ant Financial]
Deep forest is a recently proposed deep learning framework which uses tree ensembles as its building blocks. In this work, we develop the distributed version of deep forest model with XGBoost as the basic block. We present lessons and learnings of building such machine learning systems and demonstrate its effectiveness on two applications at Ant Financial; fraud detection and customer purchase behavior prediction.
Come learn more about XGBoost and how it’s utilized today!
6:00PM-6:45 PM: Doors Open
6:45PM-7:00 PM: Introduction | Speakers: Nan Zhu [Uber] and Philip Cho [Amazon]
7:00PM-7:30 PM: Machine Learning with Large-scale Telematics Data: Wayne Zhang and Gorkem Ozkaya [Uber]
7:30PM-8:00PM: Learning deep forest using XGBoost: Yin Lou [Ant Financial]
8:00PM-8:30 PM: Q&A & Networking
This is an event you don’t want to miss!
Only snacks will be provided for this event.