Termsslug: knowledge-graph

Knowledge Graph

知识图谱

A semantic network storing entities and their relationships in graph structure

Curated Intelligence

9 items · Page 1/1
Hacker News AIT2·Agents
58 pts
Agent Memory Formats Fragment as Calls Grow for a Shared Standard

Agent memory today is scattered across AGENTS.md files, /memories directories, Copilot's repo-side facts and MCP knowledge graphs, none of which can read the others, so the same lesson gets rebuilt repeatedly. The author argues the real gap is not another tool but a shared format and convention akin to OpenTelemetry, using LoreKit's plain-Markdown offline store as a concrete, inspectable example.

InfoQ 中文T2·Models
45 pts
Financial Ontology Construction: Four Elements and Three-Question Validation

This article explains how to construct business ontologies in financial data query scenarios using four core elements—concepts, attributes, relationships, and constraints—to translate business knowledge into structured language understandable by AI. It proposes a 'Three-Question Validation' method (Can AI understand? Can systems execute? Can the ontology stay fresh?) as an evaluation standard, comparing five industry construction routes while emphasizing that selection must balance business fit and engineering feasibility.

Hacker News AIT2·Tools
42 pts
Netflix Builds Ontology-Driven Observability Knowledge Graph

Netflix engineers detailed their implementation of an ontology-driven observability knowledge graph to manage massive real-time streaming logs. This approach leverages semantic modeling to correlate system dependencies and accelerate root cause analysis in environments where traditional dashboards fail.

InfoQ 中文T2·Agents
68 pts
Tencent Cloud Proposes AI Native SRE Agent: Achieving Understandable, Operable, and Verifiable Runtime via Panoramic Graphs

Tencent Cloud introduces an AI Native SRE Agent architecture to address the lag in runtime intelligence compared to coding. By leveraging a unified panoramic runtime graph as the foundation, supported by Harness mechanisms and evaluation loops, the solution enables multi-agent collaborative decision-making, making complex system states understandable, operable, and verifiable while overcoming traditional SRE toolchain fragmentation.

OpenAI 官方博客T1·Industry
42 pts
Atlassian and OpenAI Expand Partnership to Integrate GPT-6 into Enterprise Workflows

Atlassian expands its partnership with OpenAI by integrating GPT-6 frontier models into its platform and Rovo agents. Leveraging the Teamwork Graph as an enterprise context layer, the collaboration aims to automate planning, building, and delivery workflows. Additionally, Atlassian reports widespread internal adoption of Codex and plans to use DX tools to measure AI's impact on developer productivity.

腾讯技术工程 (微信公众号)T2·Agents
68 pts
Agent for SRE Error Code Troubleshooting: Minute-Level Root Cause Analysis

A team built an error code governance Agent on an orchestration platform, integrating knowledge bases, code graphs, observability data, and code hosting to reduce troubleshooting from 3-8 hours to minutes. In a 50-case comparison, the consistency of repair suggestions (action_type) with human annotations rose from 66% to 88%, with hard conflicts dropping to 0%. The core methodology involves cross-validating multi-source evidence, using workflows for structured output, and continuous iteration based on bad case evaluation.

Google AI BlogT1·Open Source
42 pts
UN Launches Open Data Platform Built on Google's Data Commons

The UN system has launched the 'UN System Data Commons,' an open, AI-ready platform built on Google's open-source project Data Commons. This initiative aims to break down data silos across UN agencies by integrating scattered global statistics into a single searchable resource, thereby accelerating data analysis and decision-making.

LangChain BlogT1·Tools
35 pts
LangChain Releases Guide on Using LangSmith for LLM Fine-tuning and Evaluation Workflows

The LangChain team released a tutorial demonstrating how to use LangSmith for dataset management and evaluation in LLM fine-tuning workflows. Using knowledge graph triple extraction as a test case, the guide shows exporting training data from LangSmith to fine-tune both open-source LLaMA2-7b-chat (via Colab/HuggingFace) and OpenAI's gpt-3.5-turbo, followed by evaluating results within LangSmith.