Building a Winning Deep Learning Recommender System - Grandmaster Series Episode 5

Dive into the fifth episode of the Grandmaster Series to learn how Kaggle Grandmasters of NVIDIA (KGMON) constructed a winning Deep Learning Recommender System for the Booking.com Data Challenge. Explore the intricacies of building a real-time recommendation system using anonymized accommodation reservation data. Gain insights into Matrix Factorization Ensemble, three distinct models (MLP, GRU, and XLNet with Session-based Matrix Factorization), and their combination into a powerful ensemble. Discover data augmentation techniques and the capabilities of NVIDIA Merlin for GPU-accelerated recommender systems. Benefit from a Q&A session and access six additional resources to deepen your understanding of advanced recommendation systems and GPU-accelerated data science.

Syllabus

– Intro to Episode Five
– Booking.com Challenge Overview & Summar
– Matrix Factorization Ensemble Overview
– Model One, MLP with Session-Based Matrix Factorization
– Model Two, GRU with MultiStage Session-based Matrix Factorization
– Model Three, XLNet with Session-based Matrix Factorization
– Combing All Three Models into One Ensemble
– Data Augmentation Approach
– NVIDIA Merlin is an End-2-End Library for GPU-Accelerated Recommender Systems
– Q&A Session