Welcome to the Semantic Search Module! We are Jay and Luis, your instructors.
In this module, you’ll learn about information retrieval systems (search), and how to incorporate them with LLMs to build powerful applications. You’ll start learning what is keyword search and dense retrieval in Chapter 2, with a small toy example. You’ll learn the important role that embeddings have in dense retrieval. For the next 3 chapters, you’ll be applying your knowledge by searching for answers to questions in a large Wikipedia dataset. In Chapter 3, you’ll use keyword search in Wikipedia, and you’ll notice that for some queries, keyword search is very successful, but for other queries it isn’t. Then in Chapter 4 you’ll use dense retrieval to search for the same queries, and you’ll notice its strength. In Chapter 5, you’ll be able to improve both keyword search and dense retrieval using a very powerful method called ReRanking. In Chapter 6, you’ll join the power of search and generative models to be able to answer questions in sentence format. And finally, in Chapter 7 you’ll learn a way to evaluate search models, and see how well they are performing.
Updated about 1 month ago