Searching video content with AI
summary
Describes extracting features to search for a specific frame across a collection of 40,000 videos using vector search; explains techniques to keep the vector database from growing excessively large. Covers using an English multimodal model to understand Russian input (methods to make the model handle Russian).
AI-written summary · about
chapters (10)
- 0:00Introduction
- 2:14Search engine challenges
- 5:49Scene search problem
- 10:37Whisper speech recognition
- 17:30Non-verbal audio event search
- 21:09How video search works
- 26:29Complex multi-database queries
- 31:35Pipeline optimization
- 33:24Infrastructure and scaling
- 35:09Q&A
related talks
33mArchitecting Scalable Multi-Modal Video Search
57mUsing Vector Databases for Multimodal Embeddings and Search
59mContextual search with vector search: exploring your options with open source tools
35mLarge-Scale Video Intelligence
48mEnter the Brave New World of GenAI with Vector Search
40mVector Search Made Simple: Getting Started with OpenSearch for AI Applications
13mThe State of AI Powered Search and Retrieval
46mNatural Language & Search, Improve your Search w/ Vector capabilities