Python + AI: Understanding Vector Embeddings

In our second session of the Python + AI series, we'll dive into a different kind of model: the vector embedding model.

A vector embedding is a way to encode a text or image as an array of floating point numbers. Vector embeddings make it possible to perform similarity search on many kinds of content.

In this session, we'll explore different vector embedding models, like the OpenAI text-embedding-3 series, with both visualizations and Python code. We'll compare distance metrics, use quantization to reduce vector size, and try out multimodal embedding models.

If you'd like to follow along with the live examples, make sure you've got a GitHub account.

📌 This session is a part of a series. Learn more here: https://aka.ms/PythonAI/2
Explore the slides and episode resources: https://aka.ms/pythonai/resources
Check out the demos: https://aka.ms/python-openai-demos

Chapters:
00:08 – Welcome & Housekeeping
01:03 – Introduction to Vector Embeddings
02:24 – Why Vector Embeddings Matter
03:32 – How Embedding Models Work
06:01 – Comparing Embedding Models
10:55 – Generating Embeddings with OpenAI
20:59 – Understanding Similarity Spaces
24:47 – Cosine Similarity Explained
34:02 – Vector Search with Exhaustive Search
40:01 – Approximate Nearest Neighbor (ANN) Search
46:54 – Compressing Embeddings: Quantization
56:17 – Compressing Embeddings: Dimensionality Reduction
59:57 – Oversampling for High-Quality Retrieval
1:00:08 – Wrap-Up & Resources

#MicrosoftReactor #learnconnectbuild

[eventID:26293]