Using Machine Learning to Personalize Web Experiences
How might we personalize web content experiences based on subtle elements of a person’s digital persona?
Standard personalization mechanisms recommend content based on a user’s profile or past activity. Search for a shoe? Here are some more shoes. Read about celebrities? Maybe you would like this other article about celebrities! We can do better.
Deep-Feeling is a proof-of-concept project that uses machine learning techniques to make better content recommendations to users. Using the Instagram API to access a user’s stream-of-consciousness, we’re filtering their feeds through a computer-vision API and using it to detect and learn subtle themes about the user's preferences.
Once we have an idea of the kind of experiences the user thinks are worth sharing, we then match the user’s characteristics against our own databases. In this proof-of-concept, we’re recommending travel experiences based on the kinds of things a person shares, using Acquia lift and Drupal 8. The applications of deep learning for content recommendation, however, are wide reaching and pretty awesome.
This is an intro session, no previous Drupal or machine learning background is required.
What you’ll learn:
- How you can use machine learning to surprise and delight users
- How a machine can make better recommendations than you could program yourself
- How to apply emerging technologies to your current and future design challenges