# Relational manifold learning with Geometric Transformers

- **Registry ID:** `catagi-gt-manifold`
- **Book:** Categories for AGI
- **Documentation status:** `public-code-and-results-linked`
- **Experimental status:** controlled synthetic recovery and registered real-data calibration
- **Canonical documentation URL:** https://categorical-ai.sridharmahadevan.com/experiments/catagi-gt-manifold

## Purpose

Triangle-aware relational recovery, dimension diagnostics, Hodge controls, and the registered MovieLens negative result.

## Book location

- Chapter 12: Manifold Learning with Geometric Transformers

## Documented studies

- Controlled relational recovery
- Post-confirmatory dimension diagnostic
- Pooled Hodge baseline
- MovieLens registered negative calibration

## Associated code packages

- `lincs-radar` — [LINCS RADAR package](https://github.com/sridharmahadevan/LINCS/tree/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar); **public**; Apache-2.0. Relational manifold, descent, chart-repair, and MovieLens studies.

## Start here

These concrete files are selected from the complete resolved source surface. They orient the reader; they are not a claim that every family is independently reproducible.

- **implementation · public-pinned** — [`lincs-radar:relational_descent.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/relational_descent.py). Open the pinned public source file.
- **runner · public-pinned** — [`lincs-radar:run_relational_dimension_diagnostic.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_relational_dimension_diagnostic.py). Open the pinned public source file.
- **runner · public-pinned** — [`lincs-radar:run_hodge_baselines.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_hodge_baselines.py). Open the pinned public source file.
- **runner · public-pinned** — [`lincs-radar:run_movielens_experiment.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_movielens_experiment.py). Open the pinned public source file.

## Entry points

- `relational_descent.py`
- `run_relational_dimension_diagnostic.py`
- `run_hodge_baselines.py`
- `run_movielens_experiment.py`

## Result artifacts

- `RELATIONAL_CONFIRMATORY_RESULTS.md`
- `RELATIONAL_DIMENSION_DIAGNOSTIC.md`
- `HODGE_BASELINE_RESULTS.md`
- `MOVIELENS_RESULTS.md`

## Resolved code surfaces (12)

- **result-record** — [`lincs-radar:HODGE_BASELINE_RESULTS.md`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/HODGE_BASELINE_RESULTS.md)
- **result-record** — [`lincs-radar:MOVIELENS_RESULTS.md`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/MOVIELENS_RESULTS.md)
- **result-record** — [`lincs-radar:RELATIONAL_CONFIRMATORY_RESULTS.md`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/RELATIONAL_CONFIRMATORY_RESULTS.md)
- **result-record** — [`lincs-radar:RELATIONAL_DIMENSION_DIAGNOSTIC.md`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/RELATIONAL_DIMENSION_DIAGNOSTIC.md)
- **runner** — [`lincs-radar:analyze_movielens_results.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/analyze_movielens_results.py)
- **runner** — [`lincs-radar:analyze_relational_confirmatory.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/analyze_relational_confirmatory.py)
- **implementation** — [`lincs-radar:hodge_baselines.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/hodge_baselines.py)
- **implementation** — [`lincs-radar:movielens_experiment.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/movielens_experiment.py)
- **implementation** — [`lincs-radar:relational_descent.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/relational_descent.py)
- **runner** — [`lincs-radar:run_hodge_baselines.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_hodge_baselines.py)
- **runner** — [`lincs-radar:run_movielens_experiment.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_movielens_experiment.py)
- **runner** — [`lincs-radar:run_relational_dimension_diagnostic.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/radar/run_relational_dimension_diagnostic.py)

## Frozen run records (0)

- No frozen run-level record is registered. Consult the result-artifact list and evidence boundary above for the exact surviving evidence; absence of a run record is not evidence that no experiment was run.

## Evidence boundary

The synthetic advantage is specific to the declared two-dimensional bottleneck, and the MovieLens study is a retained negative result rather than evidence of universal relational benefit.

A code or artifact link establishes traceability. It does not by itself establish independent reproduction, statistical adequacy, correctness, or support for a claim beyond this boundary.