Feb 24
Loading video player...
Free Course + Certificate - https://www.analyticsvidhya.com/courses/?utm_source=yt_av&utm_medium=video Timestamps Part 1: The Basics 0:00 - Introduction to Josefa & Course Roadmap 0:50 - Why Vector Databases Matter for AI (ChatGPT & Beyond) 1:45 - What is Semantic Search? (Meaning vs. Keywords) 2:31 - Traditional Databases vs. Vector Databases (MySQL vs. FAISS) 4:42 - What You Will Learn & Prerequisites 6:55 - Core Features: How Vectors Represent Real-World Data 8:54 - Mapping Meaning: The Latitude/Longitude Analogy 11:46 - Metadata Filtering in Vector DBs 14:05 - Types of Vector Databases (Open Source, Cloud, Hybrid) Part 2: Embeddings & Visualization 19:33 - Understanding Embeddings: Giving Meaning to Words 22:45 - Word Embeddings vs. Sentence Embeddings 24:25 - How Neural Networks Predict Vectors 25:56 - Visualizing High-Dimensional Data (The PCA & T-SNE Hack) 27:24 - Real-World Applications: Clustering & Recommendation Systems 32:05 - Hands-on Python: Generating Embeddings with Hugging Face 35:00 - Hands-on Python: Calculating Similarity Scores in Code 38:32 - Hands-on Python: Visualizing Clusters with Matplotlib Part 3: The Mathematics of Similarity 41:16 - The Math Behind Similarity Search 42:51 - Cosine Similarity Explained (Calculating the Angle) 46:58 - Dot Product: When Magnitude Matters 48:06 - Euclidean Distance: The Geometric Approach 49:38 - Comparison: Which Metric Should You Use? Part 4: FAISS (Facebook AI Similarity Search) 57:36 - Introduction to FAISS (Meta's Powerful Engine) 59:21 - Installation Guide: CPU vs. GPU (NVIDIA/CUDA) 1:00:28 - 4 Steps to Building a Vector Database 1:01:58 - Indexing Algorithms: Flat Index vs. IVF 1:04:36 - HNSW: The Graph-Based Speed Demon 1:05:40 - Product Quantization (PQ): Saving Memory 1:06:50 - Key Takeaways: Choosing the Right Index