Tag: #graph-database6 results
GitLab hosted a hackathon where developers used GitLab Orbit, a queryable graph of code and infrastructure, to build advanced AI agents. Participants created tools for automated impact analysis, codebase comprehension, and root-cause diagnosis by leveraging the Model Context Protocol.
BaryGraph introduces a novel knowledge graph architecture where relationships are first-class documents with their own vectors, rather than simple edges. This approach enables the discovery of structural bridges between concepts that lack semantic similarity in standard embedding spaces.
This article explores the limitations of standard vector-only RAG in enterprise environments and proposes a hybrid 'Graph RAG' architecture. It emphasizes the importance of maintaining data topology and structural relationships during ingestion to improve LLM reasoning accuracy.
Larql decomposes static LLM models into a graph database, enabling KNN walks on layers equivalent to matrix multiplication. It allows updating internal factual knowledge by inserting data into the DB without retraining. The tool uses less memory and was created by IBM's CTO.
Kumiho introduces a graph-native cognitive memory architecture for AI agents, grounded in formal AGM belief revision semantics. It unifies cognitive memory and versioned assets using primitives like immutable revisions and URI addressing. Achieves SOTA results on LoCoMo (0.565 F1) and LoCoMo-Plus (93.3% accuracy), outperforming Gemini.
Seattle startup Vieu has launched its "Business Graph," an AI-ready mapping tool that tracks professional relationships through observed data like board affiliations and work history. It aims to provide more accurate relationship intelligence than self-reported data found on platforms like LinkedIn.