This document discusses machine learning techniques for powering personalized recommendations on streaming services like Tubi.
It describes how Tubi uses over 70 models across various recommendation tasks like content and container ranking, search, and cold starting. Embedding models are used extensively to understand content and generate user representations from watch histories.
The document then focuses on Tubi's journey developing its retrieval system, moving from popularity-based filtering to personalized collaborative filtering models to current multi-interest embedding approaches that cluster users' watch histories and sample content based on cluster importance.
27. 1st & 3rd
Party Data
Audience
Assessment
Viewer-oriented data
Title-oriented
data
Products
Models
Embeddings (CTXT, MD, MMD,
Genre, Demos, Actor, et al)
Universe of Content + Metadata
Use Cases
Beam from
Universe to
Tubiverse
Cold➔
Warm➔Hot
Starting
Content Value
Assessment
Tiering
Inventory in
Tubiverse
Augmented
Search
Seeding
Growth
Coordinated
Pursuit of New
Audience
Portfolio
Analysis /
Simulation
Spock Platform
29. Overview of ML for AdTech
29
Audience Segments: Leverage
data to generate Audience segments
for targeting Ad break finder: Detect where to place an Ad break
in a video using Computer Vision
Time series forecasting: Forecast Ad Opportunities
Ad Understanding: Understand what an ad is about.