epistemic-edge 0.0.1

Air-gapped neuro-symbolic AIoT framework: 1-bit LLM cognition over mathematically verified, temporally valid edge state.

epistemic-edge

Air-gapped, neuro-symbolic AIoT framework: 1-bit LLM cognition over mathematically verified, temporally valid edge state.

Architecture

Epistemic Edge orchestrates four tiers into a strict verify-decay-generate pipeline:

Tier	Layer	Engine	Function
1	Transport	cbor-ld-ex	Hyper-compressed binary payloads over MQTT/CoAP
2	Trust	jsonld-ex	Subjective Logic fusion + PROV-O audit trail
3	Memory	chronofy	Temporal-Logical Decay Architecture (TLDA)
4	Cognition	llama-cpp-python	Grammar-constrained 1-bit LLM inference

The key insight: 1-bit quantized models lose continuous probabilistic nuance, making them susceptible to hallucination over conflicting or stale context. Epistemic Edge solves this by mathematically guaranteeing the truth of the state graph before the LLM ever touches it.

Installation

# Core framework (transport + trust + memory)
pip install epistemic-edge

# With local LLM inference
pip install epistemic-edge[llm]

# With MQTT/CoAP transport
pip install epistemic-edge[transport]

# Everything
pip install epistemic-edge[all]

Quick Start

import asyncio
from epistemic_edge import EdgeNode
from epistemic_edge.memory import DecayConfig

async def main():
    node = EdgeNode(
        node_id="gateway_alpha",
        llm_path="./models/bonsai-8b-1bit.gguf",
        decay=DecayConfig(mean_reversion_rate=1.5, threshold=0.2),
    )

    @node.guardrail(action="close_valve")
    def check_safety(state, intent):
        return state.max_uncertainty() < 0.15

    @node.on_actuate
    async def execute(intent, receipt):
        print(f"Executing: {intent.action} on {intent.target}")
        print(f"Audit trail: {receipt}")

    await node.start()

asyncio.run(main())

Core Libraries

• jsonld-ex - JSON-LD 1.2 extensions with Subjective Logic, FHIR R4, PROV-O
• cbor-ld-ex - Compact Binary Linked Data for constrained IoT networks
• chronofy - Temporal validity framework implementing TLDA

License

MIT