01
Problem and philosophy
Origin expands the question from how to mutate an equation to what objects should describe the problem. Trajectories, graphs, manifolds and stochastic processes enable different equations and conjectures. Domain paths and mathematical and physical ontologies become search spaces.
The design constructs mappings before equations. NovelMapping records physical and mathematical objects, property correspondences and forward and reverse maps. Source paths and historical mapping neighbors belong to the candidate’s identity.
02
From input to candidate
Domain Resolver 2.0 extracts a field and uses the existing Forge discovery layer for tensions. Six domain, seven mathematical and seven physical operators compose D/M/P paths, which are deduplicated and clustered into mappings.
Mappings are compared with twelve built-in mapping seeds and twelve prior traces, then generate ASTs and numerical reports. Deterministic CoMath executors produce outputs and review records, followed by specific ideas, kill tests and persistent path/failure memory.
03
Modules and state
Explore objects and mappings before deriving equation representations. Search paths, historical derivation lineages and falsification tests belong to the candidate record.
Mapping statuses distinguish unverified, known reuse, a possible ontology shift, triviality and invalidity. They expose relations to a built-in history library. A possible-new label must be interpreted through that library and method, not substituted for literature validation.
Path and failed-pattern memory preserve more than successful objects. They support re-entry through failures, explored combinations and mapping neighbors, retaining context without guaranteeing improved scientific results.
- Domain and tension
Domain Resolver 2.0 extracts a field and runs the underlying Forge discovery layer for tensions.
- D/M/P search
Six domain, seven math and seven physics operators compose, deduplicate and cluster paths into NovelMapping objects.
- Known mapping library
Twelve mapping seeds and twelve prior traces label known reuse versus candidate ontology shifts.
- AST and outputs
Mappings generate equation ASTs, small probes and deterministic CoMath outputs; specific ideas include kill tests and persistent path memory.
04
Inputs, outputs and contracts
The mapping schema requires source D/M/P IDs, objects, properties, an ontology shift, forward and reverse maps and status. Known reuse and a possible new representational shift can be labeled separately rather than collapsed into one score.
Outputs include ontology paths, objects, mappings and traces, ASTs, simulation reports, CoMath outcomes, reviewer verdicts and failed-pattern memory. Formal-looking property and evidence fields still require provenance checks.
- Inputs
Direction, mode, tensions, D/M/P operator libraries and known mapping/prior-trace libraries.
- Outputs
Ontology path/object/mapping/trace JSONL, ASTs, simulation reports, CoMath outcomes, reviewer verdicts, specific ideas and failed-pattern memory.
05
A conceptual reading scenario
For a constrained-flow question, compare adding terms in the original variables with using tangent-space or graph representations. Inspect each mapping and resulting AST. This illustrates representation choice; the contemporary heat probe still executes fixed dynamics, not the AST merely because its ID is recorded.
06
Changes from the preceding stage
Compared with Forge’s genome repairs, Origin adds first-class mappings and D/M/P search, concrete edits for four repair operators and deterministic CoMath output functions. Changes span representation, generation and execution interfaces.
Genesis’s later audit identifies category-derived placeholder properties and probes that ignore AST terms. Origin creates a richer representation space; compatibility, quality filtering and equation-sensitive execution remain subsequent improvements.
07
What the source establishes
Source identifies operators, path memory, mapping-to-AST construction and deterministic executors. Novelty labels derive mainly from a small built-in library; fixed conjectures and heat, decay, graph and stochastic probes limit claims. V5.6 ideas do not independently traverse a complete Aurora chain.
First-class mappings, Python D/M/P operators, path memory, mapping-to-AST construction, concrete edits for four repair operators and deterministic CoMath output functions.
08
Historical scope and next steps
Contemporary probes hardcode heat/decay/graph/stochastic dynamics and mostly use AST IDs; generated equations are not generally solved. Mapping novelty labels reflect a small built-in library, not external prior-art validation; v5.6 ideas do not independently traverse a complete Aurora chain. CoMath outputs are deterministic and template-driven, without machine proof, real agent research or human expert validation.
Reading example · Explaining the workflow
Reading example: trace one candidate
For a constrained-flow question, compare adding terms in the original variables with using tangent-space or graph representations. Inspect each mapping and resulting AST. This illustrates representation choice; the contemporary heat probe still executes fixed dynamics, not the AST merely because its ID is recorded.
- 01
Compare original-variable and new-object paths from a tension.
- 02
Inspect the mapping’s objects, properties and maps.
- 03
Trace the mapping to its AST and numerical report.
- 04
Check what the probe computes, then read kill tests and failure memory.
Locate the saved candidate, diagnostics and status, and identify evidence that remains absent.
This is a conceptual reading scenario derived from archive structures, not a new run, discovery result or performance experiment.
Changes in this version
Search unit
- Earlier approach
Equation genomes and repairs
- This version’s approach
D/M/P objects and mapping paths
Explore representational choices.
Repair semantics
- Earlier approach
Special operators can fall through
- This version’s approach
Four operators gain concrete edits
Connect routing to actual edits.
CoMath outputs
- Earlier approach
Primarily goals and states
- This version’s approach
Deterministic executors produce content
Supply artifacts for later checks.