Velocity Versus Volume (Expedia) Data Platforms 2018

What You’ll Learn:

1. Introduction to Bandit Models:
   • Discover the principles of bandit models, a novel approach to balancing exploration and exploitation in recommendation systems.
2. Closing the Data Loop:
   • Learn how Expedia navigates the complexities of streaming data to collect meaningful insights and refine their models for better predictions.
3. Combating Bias in Recommendations:
   • Understand the challenges of introducing new options (like hotels) into the recommendation system and the strategies employed to ensure a diverse, unbiased selection.
4. Implementation of Bandit Models:
   • See how Expedia applies bandit models to select and rank new hotels, balancing the need for data on new options with the efficiency of proven recommendations.
5. Stochastic Sorting for Better Recommendations:
   • Explore the stochastic sorting technique, which introduces a bias towards more favorable options while allowing room for exploration.
6. Feedback and Iteration:
   • Gain insights into the feedback mechanisms that allow for continuous improvement of the recommendation system, ensuring that it adapts over time to user preferences and behaviors.
7. Future Directions and Challenges:
   • Delve into the potential future applications of bandit models and the ongoing challenges in scaling and refining recommendation systems.