{
  "id": "topic:dense-retrieval",
  "_kind": "Topic",
  "_file": "domain/topics/topics-knowledge-patterns.yaml",
  "_cluster": "domain",
  "attributes": {
    "displayName": "Dense Retrieval",
    "description": "Dense Retrieval as a cross-cutting topic — similarity search using\ndense vector representations of queries and documents. Covers\napproximate nearest neighbor algorithms (HNSW, IVF, DiskANN),\nembedding space geometry, query-document asymmetry (bi-encoder\narchitectures), distance metrics (cosine, dot product, L2), and\nthe fundamental trade-off between index build time, memory usage,\nand retrieval latency. Dense retrieval excels at semantic matching\nbut struggles with exact keyword lookups and rare terms.\n"
  },
  "outgoingEdges": [
    {
      "from": "topic:dense-retrieval",
      "to": "domain:software-engineering",
      "kind": "applies_to"
    },
    {
      "from": "topic:dense-retrieval",
      "to": "domain:data-science",
      "kind": "applies_to"
    }
  ],
  "incomingEdges": [
    {
      "from": "domain:knowledge-management",
      "to": "topic:dense-retrieval",
      "kind": "contains"
    },
    {
      "from": "tool:chromadb",
      "to": "topic:dense-retrieval",
      "kind": "relates_to_topic",
      "attributes": {}
    },
    {
      "from": "tool:weaviate",
      "to": "topic:dense-retrieval",
      "kind": "relates_to_topic",
      "attributes": {}
    },
    {
      "from": "tool:pinecone",
      "to": "topic:dense-retrieval",
      "kind": "relates_to_topic",
      "attributes": {}
    },
    {
      "from": "tool:qdrant",
      "to": "topic:dense-retrieval",
      "kind": "relates_to_topic",
      "attributes": {}
    },
    {
      "from": "tool:milvus",
      "to": "topic:dense-retrieval",
      "kind": "relates_to_topic",
      "attributes": {}
    },
    {
      "from": "topic:rag-pipeline-design",
      "to": "topic:dense-retrieval",
      "kind": "related_topics"
    }
  ]
}