Collaborative Filtering (CF) is a technology commonly used in recommendation systems to suggest items or predict users' interests by collecting preferences or tastes from many users. It is based on the assumption that people who agreed in the past will agree in the future, and they will like similar kinds of items.
How Collaborative Filtering works
Collaborative filtering works in two main ways: user-based and item-based.
In user-based CF, the algorithm identifies users similar to the active users based on their previous ratings. It then predicts the interest of the active user for an item not yet rated, by calculating weighted averages of the ratings of these 'similar' users for the same item.
In item-based CF, instead of finding user's look-alike, it concentrates on mapping the relationship between the items and rating patterns. The prediction is made by first identifying items that are similar to the items the user has rated and then generating ratings for the items based on these identified similar items. The similarity among items is determined based on the ratings given by users who have rated both items.
Collaborative filtering can be a powerful tool for recommendation systems, but there are some common challenges such as cold start (no data to start with for a new user or item), scalability (handling a large number of users or items), and sparsity (the number of items greatly outnumbers the number of ratings from an individual user).
Download this guide to delve into the most common LLM security risks and ways to mitigate them.
Lakera Guard protects your LLM applications from cybersecurity risks with a single line of code. Get started in minutes. Become stronger every day.
Several people are typing about AI/ML security. Come join us and 1000+ others in a chat that’s thoroughly SFW.