Introduction to Embeddings with the OpenAI API

Enable Powerful AI Applications

Embeddings allow us to represent text numerically, capturing the context and intent behind the text. You'll learn about how these abilities can enable semantic search engines, that can search based on meaning, more relevant recommendation engines, and perform classification tasks like sentiment analysis.

Create Embeddings Using the OpenAI API

The OpenAI API not only has endpoints for accessing its GPT and Whisper models, but also for models for creating embeddings from text inputs. You'll create embeddings using OpenAI's state-of-the-art embeddings models to capture the semantic meaning of text.

Build Semantic Search and Recommendation Engines

Traditional search engines relied on keyword matching to return the most relevant results to users, but more modern techniques use embeddings, as they can capture the semantic meaning of the text. You'll learn to create a semantic search engine for a online retail platform using OpenAI's embeddings model, so users can more easily find the most relevant products. You'll also learn how to create a product recommendation system, which are built on the same principles as semantic search.

Utilize Vector Databases

AI applications in production that rely on embeddings often use a vector database to store and query the embedded text in a more efficient and reproducible way. In this course, you’ll learn to use ChromaDB, an open-source, self-managed vector database solution, to create and store embeddings on your local system.