This course offers an in-depth exploration of text retrieval using text embeddings, a combination often referred to as semantic search. It covers the theory behind text embeddings, the process of generating embeddings with both traditional and state-of-the-art models, and how to leverage these embeddings to build a text retrieval system. Additionally, it explores how to scale these systems using high-performance vector databases.
A brief introduction to the course and the problem space of text retrieval. We'll define the problem, its history of solutions, and why it's so important to the field of NLP.
A deep-dive into the intuition and theory behind text embeddings. We'll take a look at how embeddings work, the original text embedding models, and how to build semantic search pipelines.
A guide to generating powerful contextual text embeddings for semantic search using the breakthrough BERT model.
A primer on how to deploy OpenAI's Embedding API to generate state-of-the-art embeddings for text retrieval and improve the performance of semantic search pipelines.
An in-depth introduction to using vector databases like ChromaDB to scale up semantic search pipelines.