Active campaigns
| Campaign | Recent progress | Next steps |
|---|---|---|
| HST SLaM base backend parity | Project and Cortex ledgers retain the 24 September rerun submissions; published rows require coverage reconciliation. | Reconcile all seeds and stages with current project rows and Cortex before proposing another run. |
| HST SLaM Delaunay variant | Delaunay is a separate target/recipe; pre-speedup runs are archived, reruns recorded on 24 September. | Retrieve or account for every rerun, failed seed and missing stage; preserve the archived comparison context. |
| Point-source sampler admission | 2 October project record reports five Nautilus seeds; no gradient-sampler admission measurement. | Assess a bounded gradient-sampler admission proposal with the human; no new job is authorized. |
| Graphical and EP inference coordination | Mind campaign map and two scoping contracts preserved verbatim; implementation work remains in Mind. | Reconcile current Cortex observations and shipped fixes, then select one bounded phase. |
| Analytic Gaussian ensemble and control calibration | Cortex records accepted wave-1 baseline and 30 September independent-review caveats. | Use Cortex planned diagnostic on passing, edge and worst seeds; retain the human conclusion there. |
| N=50 NUTS versus three EP repeats | Cortex records failed wave 1 (EMFILE/OOM), wave-2 preparation and no submission. | Present the existing wave-2 proposal with current dependency evidence for decision. |
| Hierarchical slope recovery at scale | Cortex retains graphical baseline, post-processing OOM and stale EP hierarchical factor; Mind records later moments fix. | Verify the moments fix release and project mirror; recover missing summaries from existing samples before a proposed rerun. |
| IC50 EP and graphical scale ladder | Cortex records N=5/10/25 results, failed N=50 EP rung and width-interpretation caveat. | Check implementation close-outs, then propose a seeded ladder with appropriate output and mean-field diagnostics. |
Active tasks
| Task | Campaign | Status | Priority | Next steps |
|---|---|---|---|---|
| Drop the 1e-3 prior centring for shear / multipole / ell_comps in the benchmark | slam-hst-base | blocked | normal | Verify the release containing Galaxy#634 and Lens#754 before changing priors; retain zero-centre gradient witness. |
| EP campaign — phase map for the 2026 Q3 graphical/EP push | graphical-ep | needs-decision | high | Reconcile current Cortex records and human gates; this campaign map is never issued or bulk scheduled. |
| Expectation Propagation Scale-Up — Scoping | graphical-ep | needs-slicing | high | Reconcile completed wrapper work and Pulse profiling ownership before selecting a bounded remaining inference phase. |
| Graphical Model Scale-Up — Scoping | graphical-ep | needs-slicing | high | Reconcile current Cortex scale results and library dependencies; select one bounded inference phase with a witness. |
| NUTS/HMC trial on validated likelihoods | point-source-admission | needs-decision | normal | An unscoped idea, not permission to submit: reconcile point-source admission with imaging and weak-lensing candidate targets. |
Inference Results
Decision History
No decisions recorded yet.
Sampler candidates
Literature expectations, not measured PyAuto recommendations. Review a candidate and copy its investigation prompt.
dynesty
dynamic nested sampling · reviewed 2026-10-08
Expected uses
- Evidence and posterior baseline with adjustable effort allocation.
Limitations
- Prior transforms and likelihood-contour exploration need validation; no universal dimensional scaling advantage.
Integration: integrated: DynestyStatic and DynestyDynamic
Benchmark evidence: not_assessed: this literature review establishes no benchmark result
Investigation prompt
Last updated 2026-10-08 19:29:59 UTC
Active campaigns
| Campaign | Recent progress | Next steps |
|---|---|---|
| HST SLaM base backend parity | Project and Cortex ledgers retain the 24 September rerun submissions; published rows require coverage reconciliation. | Reconcile all seeds and stages with current project rows and Cortex before proposing another run. |
| HST SLaM Delaunay variant | Delaunay is a separate target/recipe; pre-speedup runs are archived, reruns recorded on 24 September. | Retrieve or account for every rerun, failed seed and missing stage; preserve the archived comparison context. |
| Point-source sampler admission | 2 October project record reports five Nautilus seeds; no gradient-sampler admission measurement. | Assess a bounded gradient-sampler admission proposal with the human; no new job is authorized. |
| Graphical and EP inference coordination | Mind campaign map and two scoping contracts preserved verbatim; implementation work remains in Mind. | Reconcile current Cortex observations and shipped fixes, then select one bounded phase. |
| Analytic Gaussian ensemble and control calibration | Cortex records accepted wave-1 baseline and 30 September independent-review caveats. | Use Cortex planned diagnostic on passing, edge and worst seeds; retain the human conclusion there. |
| N=50 NUTS versus three EP repeats | Cortex records failed wave 1 (EMFILE/OOM), wave-2 preparation and no submission. | Present the existing wave-2 proposal with current dependency evidence for decision. |
| Hierarchical slope recovery at scale | Cortex retains graphical baseline, post-processing OOM and stale EP hierarchical factor; Mind records later moments fix. | Verify the moments fix release and project mirror; recover missing summaries from existing samples before a proposed rerun. |
| IC50 EP and graphical scale ladder | Cortex records N=5/10/25 results, failed N=50 EP rung and width-interpretation caveat. | Check implementation close-outs, then propose a seeded ladder with appropriate output and mean-field diagnostics. |
| PyAutoFit search benchmark on gaussian_x3 (autofit_inference) | Wave-1 pilot stopped at 223 of 520 expected runs by the B3 wrap-up ruling (2026-10-08); pilot ranks nothing; catalogue, calibration record and inference-summary@1 published. | Run the 297 deferred wave-1 runs (and the JAX blend reference seed 1) on RAL partition ral on a PyAutoFit revision after A2, then freeze gaussian_x3@2 before any wave-2 row. |
Active tasks
| Task | Campaign | Status | Priority | Next steps |
|---|---|---|---|---|
| Drop the 1e-3 prior centring for shear / multipole / ell_comps in the benchmark | slam-hst-base | blocked | normal | Verify the release containing Galaxy#634 and Lens#754 before changing priors; retain zero-centre gradient witness. |
| EP campaign — phase map for the 2026 Q3 graphical/EP push | graphical-ep | needs-decision | high | Reconcile current Cortex records and human gates; this campaign map is never issued or bulk scheduled. |
| Expectation Propagation Scale-Up — Scoping | graphical-ep | needs-slicing | high | Reconcile completed wrapper work and Pulse profiling ownership before selecting a bounded remaining inference phase. |
| Graphical Model Scale-Up — Scoping | graphical-ep | needs-slicing | high | Reconcile current Cortex scale results and library dependencies; select one bounded inference phase with a witness. |
| NUTS/HMC trial on validated likelihoods | point-source-admission | needs-decision | normal | An unscoped idea, not permission to submit: reconcile point-source admission with imaging and weak-lensing candidate targets. |
| gaussian_x3 search benchmark — wave 1 (the pilot) | gaussian-x3-search-benchmark | blocked | normal | Pilot partial (223 of 520 runs); run the deferred cells on RAL partition ral after A2, then freeze gaussian_x3@2. |
Inference Results
autofit_inference
Integrity: ok · Freshness: freshness policy unspecified · evidence time unknown · Scientific qualification: no measurements
Captured source branch: main. Revision: ec35be493712d693b8cab4d071d97d9a242137b6. Capture time: 2026-10-08T19:29:28Z. Latest attempt: 2026-10-08T19:29:28Z.
Scope: inference. Producer revision: 91bfcead296c7b3c558db72923b6ab3cf0288931. Generated: 2026-10-08T12:57:54Z. Source commit timestamp; not a measurement or check-in time
Project dashboard · Ingest receipt
Coverage and limitations
{"excluded": [], "expected": {"reason": "No expected-run manifest yet; wave 1 (B3) defines one", "runs": null}, "observed": {"records": 0, "runs": 0}}- Verdicts are producer-asserted under protocol gaussian_x3@1, whose PLACEHOLDER thresholds are calibrated in the wave-1 pilot and then frozen
- Rows are judged only against a reference with the same dataset, backend, data seed and assertion mechanism; log evidences are compared only within one backend and one assertion mechanism
- Nested-search convergence needs the search's own termination condition: Nautilus is read from its sampler; other nested searches are not_assessed until PyAutoFit exposes it (phase A3)
- Completion is execution, not convergence; convergence is reported separately from acceptance
- Sample archives remain in project storage; current availability unknown
No measurements. An empty feed does not mean scientific success.
Current results
| Target | Stage | Sampler | Configuration | Seed | Execution | Convergence / acceptance | Sampler clock (s) | Fit call (s) | Diagnostics | Samples | Evidence |
|---|
Historical project results
| Target | Stage | Sampler | Configuration | Seed | Execution | Convergence / acceptance | Sampler clock (s) | Fit call (s) | Diagnostics | Samples | Evidence |
|---|
Comparisons and parity groups
Targets, seeds and stages are retained. No best-seed selection, cross-target ranking or automatic convergence threshold. Declared groups do not themselves certify parity.
No certified comparison pairs supplied by this producer.
Diagnostics, timing definitions and provenance
Decision History
No decisions recorded yet.
Sampler candidates
Literature expectations, not measured PyAuto recommendations. Review a candidate and copy its investigation prompt.
dynesty
dynamic nested sampling · reviewed 2026-10-08
Expected uses
- Evidence and posterior baseline with adjustable effort allocation.
Limitations
- Prior transforms and likelihood-contour exploration need validation; no universal dimensional scaling advantage.
Integration: integrated: DynestyStatic and DynestyDynamic
Benchmark evidence: not_assessed: this literature review establishes no benchmark result
Investigation prompt
Nautilus
neural-network importance nested sampling · reviewed 2026-10-08
Expected uses
- Evidence and posterior estimation for expensive gradient-free likelihoods.
Limitations
- Learned proposal coverage and importance-weight stability need checking on unseen modes.
Integration: integrated: Nautilus
Benchmark evidence: not_assessed: this literature review establishes no benchmark result