Differential Diagnosis for the Sky · Publication plan

Where this paper goes, and when

Six venues across the three communities this work sits between: multi-agent systems, automated planning, aerospace autonomy, and astronomical methods. Each has its own audience, format and clock.

Can this get in? The honest read

On the “MIT, Harvard, NASA-level” ambition

That bar is reachable, but it isn't set by formatting or venue choice. Papers at that level share four traits, and this project can have all four:

1. A clean headline result against the strongest baseline. Here that means P vs. B3 on test seeds, with the preregistration committed first. Beating B0 to B2 alone won't impress anyone.
2. Physics that domain experts trust. The current light curves are this project's own analytic approximations (see research/LOG.md). Swapping in published templates such as sncosmo and kilonova grids is what makes astronomers take it seriously.
3. A co-author from the field. An astronomy or aerospace-autonomy faculty member or JPL/NASA researcher who reads the draft. This is the single biggest lever on acceptance, and the AAMAS Blue Sky track explicitly weighs author track record.
4. A path to use. NASA adopts things that run in shadow mode next to a real alert stream (Fink, ALeRCE) and fit flight-like compute. The ESP32 testbed and the shadow-mode plan in §10 of the paper are the right seeds for this.

The credibility track: near-certain, real, and fast

No legitimate venue guarantees acceptance. Any venue that promises it for a fee is predatory, and a paper there damages credibility. These four, though, check soundness and fit rather than novelty, so careful, honest work gets in. Each leaves a permanent, citable, indexed record that this work exists and is moving.

What you can cite today

The Phase 1 existence proof. It is a toy problem (two exponential classes plus one withheld “bump” class, fixed Gaussian noise) and a single seed, so present it only as a preliminary signal in a vision or workshop paper:

Source: research/results/phase1/summary.json · commit · config

Recommended sequence

1. 20 Oct → 5 Nov → 12 Nov 2026AAMAS 2027 Blue Sky Ideas: OpenReview registration, then abstract, then the 4-page paper. This is the only venue where the work is ready as-is.
2. Nov 2026Commit PREREGISTRATION.md, calibrate β, γ and τ on validation seeds, then run Phase 2 on test seeds. Every venue below depends on this step.
3. 7 Dec → 14 Dec 2026ICAPS 2027 main track, only if Phase 2 test results exist by about 1 December. Otherwise skip it; a rushed ICAPS rejection costs more than waiting.
4. ~early 2027 · unconfirmedIWPSS 2027, if the biennial series runs. This is the space-planning home community.
5. Spring 2027 · rollingRAS Techniques & Instruments: the astronomy-facing methods paper, once real templates and the ZTF replay produce results.
6. ~1 Jul 2027 · inferredIEEE Aerospace Conference 2028 abstract, centred on the ESP32 hardware-in-the-loop work.
7. Late 2027 · rollingAIAA Journal of Aerospace Information Systems: the archival journal version of the full system, including decentralisation (H4) and the testbed.

The six venues

At a glance

Master manuscript: docs/index.html. Each venue page reuses its sections; only framing, length and emphasis change. Generated numbers: python -m research.figures.export_site_results.