#RAG
287 related articles
Ditching the Vector Database: SQLite BM25 and a 3.6ns Rust Sentinel for Multi-Agent Refactoring
An open-source hedge fund team refactored its multi-agent architecture: a factor ontology eliminates LLM polarity hallucinations, a switch from vector DB to SQLite FTS5 BM25 cuts reflective memory retrieval to 0.35ms, and a Rust SoA layout delivers a 3.6-nanosecond risk-control sentinel.
Retriever-Free Retrieval: DecisionRAG Navigates Documents with Decision Models
DecisionRAG introduces retriever-free retrieval: a decision model jointly evaluates candidate branches in a single forward pass, navigating the corpus structure layer by layer, surpassing dense retrieval with an 8B reranker without task training, in just 0.7-1.8 seconds per query.
Does RAG Still Have a Future? A 5-Year Skill Plan for AI Engineers
Starting from a Reddit community discussion, this article analyzes the future direction of RAG technology, the positioning of frameworks like LangChain, and the durable skills AI engineers should focus on building over the next 5-6 years—retrieval, evaluation, system design—offering practical career planning reference for beginners.
OntoGr: Adding Citation Traceability and Bi-Temporal Axes to Agent Memory
OntoGr Agent Memory is an open-source Python library that keeps verbatim quotes, sources, and bi-temporal axes for every fact in AI Agent memory, supporting reversible entity resolution, conflict adjudication, and four-way recall retrieval to solve Agent memory trust and traceability issues.
How Can RAG Frameworks Solve the Challenge of Detecting Implicit Misogyny?
A recent study proposes a RAG framework to detect implicit misogyny in English and Italian, integrating encyclopedic, slang, and feminist knowledge sources. Experiments show RAG improves recall on ImplicIT-Mis but not on SBIC+, while human evaluation consistently prefers RAG explanations, with feminist literature being the most stable and effective knowledge source.
Databricks AI Search Now Natively Supports Streaming Tables and Materialized Views
Databricks AI Search now natively supports Streaming Tables, Materialized Views, and Iceberg-based Managed Tables, enabling enterprises to build search and RAG apps on real-time data pipelines with lower index sync latency and operational costs.
Bilibili AI Large Model Tutorial: RAG Practical Guide from Beginner to Mastery
Bilibili creator A Shui launches a 109-episode AI large model RAG practical tutorial, from beginner to mastery, covering RAG, AIGC, multimodal and other core technologies, sharing learning paths and industrial practice experience to help learners systematically master large model technology.
LangChain from Scratch: A Deep Dive into a Hands-On RAG and Agent Tutorial
A deep dive into a Bilibili zero-foundation LangChain tutorial: analyzing the learning path around RAG and Agent scenarios, from basic calls to complete deployable projects. Suitable for students, career changers, and side-project developers.
How Can RAG Debugging Knowledge Be Retained? The Knowledge Management Pain of AI Teams
How can AI teams effectively retain and reuse knowledge from RAG debugging and audits? This article analyzes three current knowledge management modes and their pain points, and explores the technical feasibility and business challenges of an evidence-supported case memory system.
Building a Local Knowledge Base with LangChain-Chatchat + DeepSeek on Windows
This article, based on a Bilibili tutorial, details the full workflow of building a local knowledge base on Windows using LangChain-Chatchat and DeepSeek, including Anaconda environment setup, Ollama model management, mirror acceleration, and common error solutions, helping you quickly implement RAG Q&A.
Cosine Similarity Isn't Enough: rag-critic, an Adversarial Fact Auditor for LangGraph
Cosine similarity above 0.85 still produces confident hallucinations. rag-critic uses atomic proposition decomposition plus NLI entailment judgment for adversarial fact auditing in LangGraph, blocking and issuing rewrite instructions when the grounding score falls below threshold.
Chatbots, Workflows, and AI Agents: The Three Layers of AI Capability Explained
Everyday examples explain the three layers of AI capability: chatbots, AI workflows, and AI Agents—from passive answering to preset automation to goal-driven autonomous decision-making, with RAG vs. Agent clarified.
AI Agents vs. ChatGPT: The Paradigm Shift from Passive Responses to Autonomous Action
Explaining the essential differences between AI Agents and ChatGPT: from passive chatbots and preset AI workflows to LLM-driven autonomous agents, covering the ReAct framework, self-correction loops, and human-in-the-loop safeguards.
Dify × RAG in Practice: Build a Delta Force Game Assistant from Scratch
A practical tutorial on building a Delta Force game assistant with Dify and RAG: from zero-base concepts of agents and knowledge bases to local deployment, LLM integration, vector retrieval configuration, and web embedding delivery, plus a learning path toward LangChain and LangGraph.
Enterprise AI Memory Systems: The Real Cost of Bad Retrieval and Incorrect Ingestion
Analyzing the real costs of bad retrieval and incorrect ingestion in enterprise AI memory systems—from token waste to manual correction and user trust loss—with governance advice.
How Annotation Coverage Distorts RAG Retrieval Scores: An Empirical Study on Breast Lesions
An empirical study on a breast lesion RAG system shows ground truth annotation coverage significantly affects retrieval evaluation, urging clinical RAG assessments to report annotation sensitivity.
RAG Getting Started in Practice: From Principles to Building an Enterprise Knowledge Base
Bilibili RAG course highlights: from retrieval-augmented generation principles, document chunking and vectorization, to five accuracy factors and enterprise knowledge base boundaries, with fine-tuning vs RAG selection advice.
Cohere Releases Rerank 5: New Context Focus Feature Cuts Token Costs Dramatically
Cohere releases Rerank 5 reranking model, with Pro ranking first on nanoMTEB and ViDoRe V3 and Fast beating Qwen3-Reranker, plus Context Focus to cut token costs.
From Zero to AI LLMs in Three Months: A Pragmatic Self-Study Roadmap
Breaking down a Bilibili AI LLM beginner tutorial's learning path: from prompting and API calls to RAG and vector stores, then building an Agent to understand the ReAct loop. A pragmatic three-month roadmap to AI application development for career changers and productivity seekers.
yggy: An AI Tool for Building Your Own Knowledge Graph — Say Goodbye to Sourceless RAG
yggy is a schema-free knowledge graph AI tool that auto-extracts entities and relationships from documents, preserves sources, supports reversible merging, and serves teams and AI agents via API and MCP with EU hosting for data sovereignty.