Memory Lobe

Memory Lobe

Interface to semantic, episodic, and procedural memory

The Memory Lobe is Syntra Kernel’s gateway to long‑term knowledge.
It retrieves, writes, compresses, and manages memory across three major systems:

• semantic memory
• episodic memory
• procedural memory

---

🧠 Purpose

The Memory Lobe answers the question:

“What does Syntra already know that is relevant right now?”

It ensures continuity, coherence, and context.

---

🔍 Responsibilities

1. Memory Retrieval

Query semantic and episodic memory.

2. Memory Writing

Store new events, knowledge, and procedures.

3. Memory Compression

Reduce redundancy while preserving meaning.

4. Relevance Scoring

Determine which memories matter for the current task.

5. Memory‑World‑Model Synchronization

Ensure consistency between memory and the world‑model.

---

🧩 Inputs

• world‑model updates
• perception events
• planning queries
• reflection feedback
• agent requests

---

📤 Outputs

• semantic knowledge
• episodic timelines
• procedural patterns
• memory embeddings
• relevance‑ranked results

---

🔗 Interactions

• World‑Model Runtime — synchronizes entities and events
• Planning Lobe — provides knowledge for strategy generation
• Reflection Lobe — updates memory after evaluation
• ThoughtStream — logs memory access

---

🧪 Example Memory Query (SL2)

memory.query:
  type: episodic
  filter: events_related_to("checkout")
  limit: 5

---

🧭 Why It Matters

Memory is the backbone of continuity.
Without it, Syntra Kernel would be stateless and reactive — not cognitive.