Glossary

This glossary defines key terms used throughout CLE-Net documentation.

A

Agent

An autonomous CLE-Net node that observes local data, discovers rules, and participates in consensus.

Atom

A unit of meaning extracted from raw interaction (entity, action, condition, etc.).

B

Blockchain Layer

The decentralized coordination layer that stores rule commitments, enables consensus, and provides persistence.

C

CLE

Cognitive Logic Extraction. The core process of converting unstructured interaction into symbolic rules.

CLE-Net

Decentralized Cognitive Agent Network. The complete system architecture.

Consensus

The process by which multiple independent agents agree on the validity of a discovered rule.

Context

The domain, situation, or conditions under which a rule applies.

Contradiction

A signal indicating that two or more rules cannot simultaneously apply under the same context.

D

Decay

The gradual reduction in confidence of a rule that is not confirmed or is contradicted.

Discovery

The process by which an agent identifies a latent rule from observed patterns.

E

Event

A captured instance of human interaction, stored in the event stream.

Evidence

Data supporting a discovered rule.

G

Genesis State

The initial state of the network with no accepted rules.

I

Independence

The property that agents operate on distinct data without coordination.

Independence Score

A numerical measure (0-1) of how independent two or more agents are.

K

Knowledge Graph

A graph structure representing entities, relations, and rules discovered by agents.

L

Ledger

The append-only record of rule commitments and consensus outcomes.

Logic Form

The canonical symbolic representation of a rule.

M

Mining

The process of discovering rules that achieve consensus. In CLE-Net, “mining” means “thinking.”

N

Node

A participant in the CLE-Net network, running an agent.

P

PoC

Proof of Cognition. CLE-Net’s consensus mechanism based on independent discovery.

Predicate

A logical statement about entities and their attributes.

R

Rule

A symbolic representation of a discovered pattern or law.

Rule Candidate

A newly discovered rule that has not yet achieved consensus.

Rule Cluster

A group of rule commits that represent the same logical rule.

Rule Commit

A broadcast of a rule hash and metadata to the network.

S

Semantic Atomizer

The component that extracts atoms from raw interaction.

Symbol

A logical representation of meaning derived from atoms.

Symbolic Regression

The process of discovering mathematical or logical relationships from data.

Symbolic Reasoning

Reasoning using formal symbols and logic rather than statistical patterns.

Enhanced Symbolic Regression

Advanced symbolic regression using Genetic Programming (GP) for complex pattern discovery, including temporal pattern recognition and uncertainty quantification.

Genetic Programming (GP)

An evolutionary algorithm that evolves symbolic expressions to discover patterns in data.

Temporal Pattern Recognition

The identification of patterns that change over time, including trends, periodicity, and change points.

Uncertainty Quantification

Methods for estimating confidence intervals and prediction bounds for discovered rules.

Full-Duplex Interaction

Simultaneous input and output processing, allowing real-time bidirectional communication.

OCR (Optical Character Recognition)

The conversion of images or scanned documents into machine-readable text.

Graph RAG

Retrieval-Augmented Generation using knowledge graphs for context-aware responses.

T

Temporal Graph

A knowledge graph that preserves the history of changes over time.

Threshold

A configurable minimum value (e.g., confidence, independence) required for an action.