Appendix: Algorithms & Technical Details

Spikuit draws on ideas from five fields. Each page covers the theory, how Spikuit applies it, and key references.

1. Computational Neuroscience — STDP, LIF, spreading activation, sleep consolidation
2. Cognitive & Developmental Psychology — forgetting curve, testing effect, ZPD, scaffolding
3. Spaced Repetition Systems — FSRS v6, grade mapping
4. Knowledge Graphs & Graph-Based ML — PageRank, APPNP, community detection
5. Information Retrieval & RAG — hybrid retrieval, conversational curation
6. Spike-Circuit Correspondence — theoretical grounding: spikes, retrieval, and eigenthemes as three readings of one operator
7. Agent Reflection — Sibling Brain pattern: how Skills acquire self-improving memory via an isolated reflection graph
8. Emergent Structure — how domain (user-declared, nullable) and community (emergent from Synapses) coexist, and when blanks get filled automatically
9. Cognitive Retrieval & AMKB Memory Model — four-stage cognitive model (encode/store/consolidate/retrieve), R×I×R decomposition of Spikuit's score, and why sub-linear MAC is the AMKB target

For implementation specifics (algorithms, parameters, embedding pipeline):

• Implementation Details — APPNP math, STDP updates, LIF model, plasticity parameters, embedder config, tech stack

For a higher-level overview, see Concepts.