Search
Glossary3
Open the full glossary- EmbeddingTurning text into a list of numbers so that "similar meaning" becomes "close in distance".
- Vector databaseA database that searches by how close the meaning is: you give it a query vector and it returns the nearest few records, instead of matching exact wording.
- ChunkingBefore embedding, cut long documents into small pieces; a vector compresses the meaning of a whole passage into one point, and the longer the passage, the blurrier that point.
Found 7 articles
WeChat Open Source WeMM-Embedding Deep Dive: The Multimodal Embedding Model Topping MMEB-v2, Can It Run on Your Mac?
Breaking down WeMM-Embedding from the arXiv 2608.24053 technical report and source code: Qwen3.5 native multimodal backbone, <embedding> token pooling, two-stage training (Semantic-ID resampling + bidirectional KL distillation + model merging), full MMEB-v2/v3 leaderboard results, and local Mac Studio MPS benchmarks (47ms text / 321ms image / 6
Vector Memory Deep Dive: A Production-Grade Memory System That Ensures AI Agents Never Lose Memory Again
State amnesia is the #1 killer in production Agent environments. We built a four-layer vector memory system based on Qdrant + BGE-m3, with 9 retrieval modes, validated with 6,675 memories in real-world use, raising Chinese search accuracy from 0% to >78%. One line curl | bash, 100% local deployment, and data sovereignty stays in your hands.
A Lightweight Overhaul of MemoryHub in Practice: Cutting from 10 Databases to 5, and RAM from 6.2GB to 500MB
What happens when you delete Neo4j, Elasticsearch, and MongoDB from an AI memory system, keep only the Qdrant + FAISS + SQLite-vec trio, and replace the 9.1GB BGE-m3 with a 193MB bge-small-zh? A complete performance dissection, an embedding model comparison, and a trimming plan aimed at ordinary computers.
XSkill Deep Source Code Analysis: Technical Design, Risks, and Implications for MemoryHub from an ICML 2026 Paper
From the three-stage Skill lifecycle and two-layer Experience structure to cross-trajectory contrastive critique, cascading failures, embedding retrieval blind spots, and over-merging risks, this is a complete technical dissection of XSkill-Agent/XSkill, with an applicability assessment for the Agentic Research memory system.
Polyglot Persistence in Practice: A Full-Pipeline Evaluation of a Ten-Database Memory System
From BGE-m3 vector embeddings and ten-way synchronized writes to a comparison of search quality and speed across nine backends: a complete stress test of an AI memory system. Qdrant is the king of semantics, FAISS the king of speed, and Neo4j the king of relationships.
MemoryHub v2.0 Full Record of Ten-Database Sync: The 6-Hour Battle from 0 Points to 3,892 Records
A complete record of MemoryHub's technical evolution from a monolithic Qdrant to a unified embedding pipeline across ten databases: 18 bug fixes, Python scoping traps, BGE-m3 dimension errors, missing system Python dependencies, and the final architecture that achieves one embedding then ten-database sync.
RAG In-Depth Principles and Practice: A Complete Guide from Chunking to Rerank
The complete tech stack for Retrieval-Augmented Generation (RAG): Embedding selection, Chunking strategies, vector database comparison, Rerank optimization, with practical code examples.