Back to corpus
research noteexperiment writeup candidatescore 24

Graph Kernel + Kimi-K2 Memory Integration

Connect Comp-Core's Graph Kernel and RAG++ to the Kimi-K2 memory layer, enabling: - **Slice-conditioned synthesis** — Context slicing for focused responses - **Knowledge graph integration** — Semantic relationships from Graph Kernel - **Dual-plane retrieval** — Raw messages + semantic atoms

Open full HTMLDownload HTMLCorpus JSONPaper schema

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

Connect Comp-Core's Graph Kernel and RAG++ to the Kimi-K2 memory layer, enabling: - **Slice-conditioned synthesis** — Context slicing for focused responses - **Knowledge graph integration** — Semantic relationships from Graph Kernel - **Dual-plane retrieval** — Raw messages + semantic atoms ### 1. Knowledge Graph Sync Kimi-K2 extracts knowledge triples → Graph Kernel stores relationships ### 2. Slice-Conditioned Synthesis Before synthesis, get admissible context slice from Graph Kernel ### 3. RAG++ Retrieval Augmentation Use RAG++ for semantic search when building context ### Phase 1: Local Integration (Current) - [x] SQLite memory store - [x] Knowledge graph table - [ ] Graph Kernel client in Python - [ ] Slice-aware context building

Promotion decision

What has to happen next

Attach run IDs, datasets, metrics, and reproduction commands.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.