Summary

In this chapter, we explored how LLMs could change the way we approach a recommendation system task. We started from the analysis of the current strategies and algorithms to build recommendation applications, differentiating among various scenarios (collaborative-filtering, content-based, cold start…) as well as different techniques (K-Nearest Neighbors, Matrix factorization and Neural Networks).We then moved to the new, emerging research of how to apply the power of LLMs to this field, and explored the various experiments that have been done in recent months. Leveraging this evidences, we built a movies recommender applications powered by LLMs, using LangChain as AI orchestrator and Streamlit as front-end, showing how LLMs can revolutionize this fiend thanks to their reasoning capabilities as well as generalization.This was just one example of how LLMs not only can open new frontiers, but also can they enhance existing fields of research.In next chapter, we will see what...