brycewang-stanford ECAI Review Process
Model the pipeline before interpreting any single review. ECAI runs a single-round, double-blind conference review with a rebuttal — not a journal-style revise-and-resubmit. The most consequential mental shift for authors arriving from a journal (or from a venue with a major- revision decision) is that there is no second submission: the paper you submit, plus one short author response, is the whole story. The current edition, IJCAI-ECAI 2026, follows the IJCAI process because ECAI 2026 is being run jointly with IJCAI.
Process model (IJCAI-ECAI 2026)
- Submission and review are double-blind; identities are hidden from reviewers throughout.
- A summary-reject first phase filters clearly-out-of-scope or below-bar submissions early (IJCAI-ECAI 2026 summary-reject notification: 4 March 2026). A paper cleared past summary reject goes to full review.
- Full review is organized under area chairs / senior PC who oversee reviewers in a topic area and lead the discussion. Reviewers weigh significance of the AI contribution, technical soundness (proofs and/or experiments), clarity, and fit.
- Authors get one author-response (rebuttal) window (IJCAI-ECAI 2026: 7-10 April 2026) to answer factual points before the final decision.
- The decision is accept or reject (final notification 29 April 2026). There is no guaranteed revision round; camera-ready changes are minor, not a re-review.
Reading the decision categories
| Decision | What it means | Author move |
|---|---|---|
| Accept | Contribution and evidence hold | Camera-ready in the correct template; do not expand scope |
| Reject (after full review) | A soundness, significance, or clarity gap the rebuttal did not close | Revise substantively and reroute (AAAI/IJCAI/a specialist venue) or resubmit next ECAI |
| Summary reject | Out of scope, below bar, or a formatting/anonymity violation caught early | Do not appeal into the void — fix the structural issue before any resubmission |
The strategic reading: because there is no major-revision safety net, the submitted paper must already be complete — the proof finished, the baseline fair, the claim matched to the evidence. A gap you were "planning to fix in revision" has nowhere to go.
How ECAI differs from its siblings
- vs. a journal / major-revision venue: ECAI has no Major Revision decision. You cannot count on a second round; there is one rebuttal, then accept/reject.
- vs. AAAI: AAAI and ECAI are both general-AI, double-blind, single-round with rebuttal, but on
different calendars and templates (AAAI style vs
ecai.cls/ijcai.sty) and governed by different bodies (AAAI vs EurAI). Never assume shared dates or format. - vs. IJCAI (standalone years): ECAI is EurAI's European flagship; IJCAI is the international
one. In 2026 they are the same event (joint IJCAI-ECAI), so the process is IJCAI's — but in a
standalone ECAI year the FAIA/
ecai.clsconventions and PAIS co-location return. - vs. AAMAS / KR / specialist venues: those run their own reviewing for multi-agent systems or
knowledge representation; a paper matched better to a specialist reviewer pool may fare better
there than in ECAI's broad pool (
ecai-topic-selection).
Who reads you
Expect several reviewers under an area chair drawn from your topic area — and note that by submitting you volunteer for the reviewer pool yourself (IJCAI-ECAI 2026 policy). Because ECAI spans symbolic AI, planning, KR, multi-agent systems, and ML, matching your topics well is what gets you expert reviewers rather than distant ones. Vague method descriptions get caught in the discussion the area chair runs, not skimmed.
Where author leverage actually exists
[Before submission] topics/keywords -> area chair + reviewer pool (largest lever)
[Summary-reject risk] scope, formatting, anonymity, page budget correct BEFORE upload
[Rebuttal] correct factual misreadings, supply one asked-for number,
answer the specific objection each reviewer raised
[After reject] no appeal that reopens the decision; reroute or resubmit next cycle
A rebuttal moves borderline papers when it corrects a factual misreading or provides a clarification a reviewer said was missing; it does not move papers when it argues taste or promises future work. With only one response and no revision round, the rebuttal is a precision instrument, not a plan.
Reading a review packet
Weight reviews before answering. A review that cites your definitions, lemmas, and tables was read closely; its author is your likely advocate if the rebuttal holds. A review that discusses only novelty has left soundness to the others — answer each reviewer on the axis they raised. The area chair, not any single reviewer, drives the outcome in discussion, so a response that gives the AC a clean reason to champion the paper is worth more than one that "wins" against a hostile reviewer.
Misreadings to avoid
- Expecting a revision round — there is none; the submission is final bar minor camera-ready.
- Treating the rebuttal as a debate — the AC-led discussion decides; your text is evidence for an advocate, not a closing argument.
- Ignoring summary-reject risk — a formatting, anonymity, or scope slip can end the paper before review even begins.
- Projecting last year's process — a standalone ECAI and a joint IJCAI-ECAI year differ in organizer, template, and sometimes mechanics; confirm the current edition.
Output format
[Process stage] pre-submission / summary-reject phase / under review / rebuttal / final
[Edition regime] standalone ECAI / joint IJCAI-ECAI (IJCAI process)
[Decision] accept / reject / summary reject, with the criterion driving it
[Criterion map] each review point -> significance | soundness/proof | evidence | clarity | fit
[Leverage plan] the rebuttal move that can actually change this AC's recommendation
[Forbidden moves] identity leak / new unsupported claims / promising a revision that cannot happen