01
Problem and philosophy
Genesis addresses quality after Origin enlarges the combination space. More D/M/P combinations can yield identity mappings, placeholder properties or non-executable equations. Compatibility, nontriviality and supported execution become selection criteria.
Equation structure must also affect computation. If changing an AST ID still solves the same heat equation, reports cannot distinguish mechanisms. Genesis interprets supported right-hand-side terms and routes conjecture types to different proof strategies.
02
From input to candidate
Resolver 3.0 repairs zero-hit and substring behavior. A proposer interface adds grounding and anti-template filters. Compatibility graphs prune combinations before mapping construction, and quality and diversity selection limit trivial or repeated candidates.
Term dispatch computes supported diffusion, variational, projection, memory and noise contributions. CoMath routes nine conjecture types. Seven Aurora-V57 gates consume mapping, AST, simulation, prior, proof, review and idea artifacts, with persistent path memory and a blind-review queue.
03
Modules and state
More combinations do not equal better discovery. Remove placeholder properties, identity mappings and non-executable combinations before numerical and proof-route checks.
The actual V57 aggregation rejects at three failures; zero failures and at least five passes yield promising-but-unproven, while other combinations remain unverified. A failed gate or several warnings cannot be summarized as a complete pass.
Contribution and skipped-term records reveal whether AST changes affect computation. Supported semantics are finite: variational gradients use a canonical energy and projection preserves a mean. General energies, constraints and stochastic processes require additional implementations.
- Domain and proposal
Resolver 3.0 fixes zero-hit and substring errors; a grounded, anti-template-filtered D-Tree proposer interface is added.
- Compatibility and quality
Compatibility pruning precedes mapping construction, and a quality referee checks nontriviality and executable/theorem artifacts.
- Diversity selection
MAP-Elites and bucket caps limit repeated shifts, AST signatures and domain operators; prior deltas use layered path traces.
- Supported terms and conjectures
Term dispatch executes supported RHS contributions, and CoMath routes nine conjecture types to distinct strategies.
- Gates and memory
Seven Aurora-V57 gates consume candidate evidence, with cross-run path memory and a persisted blind-review queue.
04
Inputs, outputs and contracts
Selection requires source paths and executable structure, not just object categories. Numerical context includes coefficients, grids, history, constraints and graph Laplacians. Unknown terms are recorded as skipped rather than treated as implemented semantics.
Outputs add quality reports, elite mappings, type-specific proof graphs, kill tests and review cards. Cards and queues are review materials; their existence does not imply completed human expert judgment.
- Inputs
Origin paths/mappings, direction, compatibility rules, supported term evaluators, conjecture strategies and prior traces.
- Outputs
Quality reports, elites, AST/simulation reports, type-specific proof graphs and kill tests, expert-review cards/queue, Aurora results and cross-run memory.
05
A conceptual reading scenario
For a memory-bearing mapping, inspect compatibility, quality and diversity decisions, then follow the MemoryIntegral contribution and conjecture-specific proof route. Finally check that the reports refer to the same candidate.
06
Changes from the preceding stage
Compared with Origin, mappings are filtered, a finite AST vocabulary executes and proof routes vary by conjecture. Resolver, prior-delta and cross-run memory changes reduce unrelated fields and repeated paths.
Logos later finds reused V5.6 ideas, positional or default associations among artifacts and dictionary AST replay that can produce an empty RHS. Candidate linkage and serialization therefore become V5.8’s focus.
07
What the source establishes
Source confirms finite term execution, compatibility pruning, diversity selection and seven gates. The LLM-named proposer remains deterministic. Unknown-term handling and fixed semantics limit equation scope; proof routes are plans and review queues are not completed reviews.
Compatibility pruning, spurious-mapping detection, diversity selection, supported AST RHS execution, conjecture-sensitive proof routes, seven gates and a persisted review queue.
08
Historical scope and next steps
The LLM-named proposer is still a deterministic stub, and review cards/queues do not imply human endorsement. The 5.8 audit finds reused v5.6 ideas, misaligned evidence links and possible no-op replay of dictionary ASTs; candidate chains are not fully closed. Finite term solvers and proof routes are exploration prototypes, not general proof or open-world discovery systems.
Reading example · Explaining the workflow
Reading example: trace one candidate
For a memory-bearing mapping, inspect compatibility, quality and diversity decisions, then follow the MemoryIntegral contribution and conjecture-specific proof route. Finally check that the reports refer to the same candidate.
- 01
Read why compatibility and quality checks retain the candidate.
- 02
Inspect supported AST terms and skipped-term records.
- 03
Follow conjecture type to the proof route and kill test.
- 04
Cross-check identities of mapping, simulation and review card.
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
Combination quality
- Earlier approach
Generate broad ontology combinations
- This version’s approach
Prune by compatibility and quality
Reduce identity and placeholder candidates.
Equation sensitivity
- Earlier approach
Probes use preset equations
- This version’s approach
Dispatch supported RHS terms
Make structure affect computation.
Proof routes
- Earlier approach
Fixed templates and routes
- This version’s approach
Strategies for nine conjecture types
Use different routes for different questions.