Hierarchical memory, bandits, RL, knowledge tensors.
11 repos (5 original, 6 forked from Lucineer)
| Repo | Origin | Language | Description |
|---|---|---|---|
| SuperInstance/Bayesian-Multi-Armed-Bandits | ⭐ orig | TypeScript | No description |
| SuperInstance/Murmur | ⭐ orig | TypeScript | Knowledge Tensors for self-improving agents |
| SuperInstance/bandit-learner | ⭐ orig | — | Multi-armed bandit algorithms for exploration-exploitation and online learning |
| SuperInstance/causal-memory | 🍴 fork | — | Causal reasoning engine — track and query cause-effect chains across fleet actions |
| SuperInstance/experience-pipeline | 🍴 fork | — | Training data pipeline from git history — experience primitives, local LoRA, fleet federated aggregation |
| SuperInstance/forgetting-problem | 🍴 fork | — | The unsolved problem — thermal decay, scheduled pruning, and emergency memory purge |
| SuperInstance/forgiveness-function | 🍴 fork | — | Strategic forgiveness in autonomous agents — conditional momentum-based recovery |
| SuperInstance/frozen-intelligence | 🍴 fork | — | The Frozen Intelligence — mask-locked inference as a vessel class. Silicon-bound cognition for sovereign edge agents. Part of the Lucineer ecosystem. |
| SuperInstance/hierarchical-memory | ⭐ orig | Python | 6-tier memory hierarchy with consolidation, vector search, and identity persistence for AI agents |
| SuperInstance/hybrid-memory | 🍴 fork | — | Git + KV + causal memory equipment for fleet vessels |
| SuperInstance/memory-visualization | ⭐ orig | — | No description |