Research Experiments & Open References

Public Experiments

Open-source prototypes and technical benchmarks.

Focus: Prototyping & Benchmarking 12 active repositories

Overview

The experiments listed here are public-facing prototypes and research tools that support our broader focus on reliable, local, and verifiable AI. These projects explore different facets of the System 2 engineering direction, from hyperdimensional representations to constrained natural languages.

These repositories provide the empirical grounding for the theories developed in the MRP and Ruliology sections. They are open for inspection, testing, and reference.

Experiment List

  1. MRP-VM

    Virtual machine runtime for governed interpretation, explicit routing, and auditable execution.

  2. autoResearchLib

    Library for automating research workflows, literature analysis, and systematic synthesis.

  3. Executable CNL

    Executable programming language based on constrained natural language syntax.

  4. VSABrains

    Experiment suite for discrete, CPU-first learning inspired by A Thousand Brains.

  5. VSA Representations

    Experiments with VSA/HDC representations and their tradeoffs.

  6. VSAText

    Experiments on using VSA/HDC to analyze large volumes of text with partial semantic structure.

  7. VSAVM

    VSA plus symbolic virtual machine for learning and structured manipulation.

  8. SomaVM

    Learning system experiment based on a VM that simulates pleasure and pain as reward.

  9. HDC-RE

    Reasoning engine built on hyperdimensional representations.

  10. BSP

    Bitset System for Prediction: CPU-friendly continuous learning without transformers.

  11. UBHNL

    Universal Boolean Hypergraph kernel with a constrained natural language and DSL front-end.

  12. Spock

    Deterministic neuro-symbolic geometry for explainable reasoning.