{
  "id": "topic:memory-deduplication",
  "_kind": "Topic",
  "_file": "domain/topics/topics-knowledge-patterns.yaml",
  "_cluster": "domain",
  "attributes": {
    "displayName": "Memory Deduplication",
    "description": "Memory Deduplication as a cross-cutting topic — strategies for detecting\nand removing duplicate memory entries across extraction cycles. Covers\nSHA-256 content-addressable dedup (exact match, used by agentmemory),\nsemantic dedup (embedding similarity threshold, used by auto-dream\nconsolidation), and hybrid approaches that combine exact and semantic\nmatching. Without deduplication, auto-memory systems accumulate\nredundant entries that waste context window space and dilute retrieval\nquality. The key trade-off is between aggressive dedup (risking loss of\nsubtly different memories) and conservative dedup (accumulating near-\nduplicates that clutter retrieval results).\n"
  },
  "outgoingEdges": [
    {
      "from": "topic:memory-deduplication",
      "to": "domain:software-engineering",
      "kind": "applies_to"
    },
    {
      "from": "topic:memory-deduplication",
      "to": "specialization:ai-agents-conversational",
      "kind": "applies_to"
    },
    {
      "from": "topic:memory-deduplication",
      "to": "topic:auto-memory",
      "kind": "related_topics"
    },
    {
      "from": "topic:memory-deduplication",
      "to": "topic:memory-consolidation",
      "kind": "related_topics"
    }
  ],
  "incomingEdges": [
    {
      "from": "domain:knowledge-management",
      "to": "topic:memory-deduplication",
      "kind": "contains"
    },
    {
      "from": "tool:agentmemory",
      "to": "topic:memory-deduplication",
      "kind": "relates_to_topic",
      "attributes": {}
    }
  ]
}