HomepageData + AI Summit 2022 Logo
Watch on demand

Building Recommendation Systems Using Graph Neural Networks

On Demand

Type

  • Session

Format

  • Virtual

Track

  • Data Science, Machine Learning and MLOps

Industry

  • Media and Entertainment

Difficulty

  • Intermediate

Duration

  • 40 min

Overview

RECKON (RECommendation systems using KnOwledge Networks) is a machine learning project centred around improving the entities intelligence.



We represent the dataset of our site interactions as a heterogeneous graph. The nodes represent various entities in the underlying data (Users, Articles, Authors, etc.). Edges between nodes represent interactions between these entities (User u has read article v, Article u was written by author v, etc.)



RECKON uses a GNN based encoder-decoder architecture to learn representations for important entities in our data by leveraging both their individual features and the interactions between them through repeated graph convolutions.



Personalized Recommendations play an important role in improving our user's experience and retaining them. We would like to take this opportunity to walk through some of the techniques that we have incorporated in RECKON and an end-end building of this product on databricks along with the demo.

Session Speakers

Headshot of Swamy Sriharsha

Swamy Sriharsha

Data Scientist II

Condé Nast

See the best of Data+AI Summit

Watch on demand