A technical guide on implementing vector search in Oracle Database 26ai, using a car manual as a practical example to improve semantic search.
A visual essay explaining LLM internals like tokenization, embeddings, and transformer architecture in an accessible way.
Explores the application of Graph Neural Networks, embeddings, and foundation models to spatial data science, with practical examples in R.
Introducing Spelungit, a semantic search tool for Git commit history that uses natural language queries instead of exact keywords.
A guide to implementing semantic search for static websites using Amazon S3 Vector Buckets and Microsoft's Semantic Kernel for embedding generation.
Explains how LLMs work by converting words to numerical embeddings, using vector spaces for semantic understanding, and managing context windows.
An experiment comparing retrieval performance of chunked vs. non-chunked documents using long-context embedding models like BGE-M3.
Building an image search system using GPT-4 Vision and Azure AI to find images via text queries or similar pictures.
Explains how to use Azure AI Search's integrated vectorization for automatic query and field vectorization, with portal and indexer examples.
Explains AI transformers, tokens, and embeddings using a simple LEGO analogy to demystify how language models process and understand text.
Interview with Frank Liu on vector databases, embeddings, his career in ML/hardware, and work culture differences between China and the US.
Explains a chunk-based embedding method using LangChain and Pinecone to improve blog post search accuracy and efficiency.
A guide on using Redis as a vector database to store and query embeddings for semantic search, replacing Pinecone in a tech stack.
A technical guide on using Pinecone vector search and OpenAI's API to build a semantic search engine for personal blog posts.
Guide to deploying a Sentence Transformers model on Amazon SageMaker for generating document embeddings using Hugging Face's Inference Toolkit.
Explains how Graph Neural Networks and node2vec use graph structure and random walks to generate embeddings for machine learning tasks.
Explores word morphing using word2vec embeddings and A* search to find semantic paths between words, like 'tooth' to 'light'.
