AI Agent 时序记忆图谱设计模板

You are an expert AI memory system architect. Design a temporal knowledge graph-based memory system for an AI agent with the following specifications:

Requirements

• The agent needs to maintain a context graph that tracks how facts change over time
• Each fact/relationship should have a validity window (valid_from, valid_until)
• Entities should have evolving summaries updated as new information arrives
• All derived facts must trace back to source episodes (provenance)

Design the following:

1. Entity Schema: Define node types (Person, Organization, Concept, Event) with temporal metadata
2. Relationship Schema: Define edge types as triplets (Entity → Relationship → Entity) with validity windows
3. Episode Ingestion Pipeline: How raw data (conversations, documents, events) gets processed into graph updates
4. Conflict Resolution: Strategy for handling contradictory facts (e.g., user changed job)
5. Retrieval Strategy: Hybrid retrieval combining semantic search, keyword matching, and graph traversal
6. Query Examples: Show how to query "What was true at time T?" vs "What is true now?"

Output Format

Provide the design as a structured document with code examples in Python, using Neo4j or similar graph DB as the backend. Include sample Pydantic models for the ontology.