Finding out what to watch is tough choice mainly because of the abundance of options and the scarcity of time. Let’s follow a data-driven approach to arrive at a decision! In our dataset of movies, we know the
  • description,
  • genre
  • title
  • and the release_year of the movie.
Putting together a vector index on top of these embedded inputs will create a space where we can search semantically to find our movie choice for tonight. We are going to browse the movies
  • searching with an idea (heartfelt romantic comedy)
  • tweak around the results giving more imporance to matches in certain input fields
  • search in description, genre and title with different search terms for each
  • and after finding a close enough movie (that is not quite it), search around utilizing that one, too

​
Follow along in this Colab

Semantic Search - Movies Notebook

Open the interactive Colab notebook to follow along with the semantic search movies tutorial.