User Behavior Hashing for Audience Expansion - Deep Learning Approach

Explore a deep learning-based architecture for learning binary hashing of sequential behavior data in this 27-minute video from Databricks. Discover how to capture evolving user preferences and exploit activity patterns at different time scales using the Spark platform for large-scale data preprocessing, modeling, and inference. Learn about the Audience Intelligence Platform, lookalike modeling, and category models. Gain insights into Python UDFs and Python Group Map functions, and see how distributed inference jobs are performed on Databricks using Pandas UDF. Understand the application of this technology in user modeling and large-scale data retrieval scenarios, and examine the results of implementing this approach.

Introduction
Audience Intelligence Platform
Lookalike Modeling
Category Model
Python UDF
Python Group Map
Results