Ecai submission

ECAI Submission

Run this audit before uploading to the ECAI submission site. First settle which regime the target edition is in, because it changes the template, the publisher, and the page budget:

• A standalone ECAI (e.g. ECAI 2023/2024) publishes in IOS Press FAIA (Frontiers in Artificial Intelligence and Applications), uses the ecai.cls template, and allows 7 pages of body + 1 page of references.
• ECAI 2026 is the joint IJCAI-ECAI 2026 (Bremen, 15-21 Aug 2026). It follows IJCAI: the ijcai.sty template, 7 pages body + 2 pages references plus an optional ethics statement, and IJCAI's own open-access proceedings — not FAIA this cycle.

Every number below was read from the IJCAI-ECAI 2026 and ECAI 2024 calls on 2026-07-09 via search renderings (see resources/official-source-map.md); treat them as a one-cycle snapshot and reopen the live call first.

The two-deadline flow

ECAI separates an abstract/registration deadline from the paper deadline, and both are AoE:

• Abstract locks in title, abstract text, authors, and topics about a week early (IJCAI-ECAI 2026: 12 January 2026; ECAI 2024: registration 19 April 2024). Miss it and the system will not accept the PDF later.
• Paper uploads the anonymized PDF (IJCAI-ECAI 2026: 19 January 2026; ECAI 2024: 25 April 2024).

Register with the real title and abstract (plain text, ~100-300 words), not a placeholder — the abstract drives reviewer bidding and area-chair assignment, and a registered abstract that diverges from the final paper quietly worsens your match.

Format and page budget

• Template: ijcai.sty for IJCAI-ECAI 2026; ecai.cls for a standalone ECAI. This is not the ACM acmart path, not IEEE IEEEtran, and not LNCS — do not carry a sibling-venue habit over.
• Page budget (verify per cycle): 7 pages of body for all text and figures. References are the only overflow — 1 page (standalone ECAI) or 2 pages (IJCAI-ECAI 2026). There is no unlimited appendix inside the paper; anything a reviewer must read to judge the paper lives in the 7-page body.
• Ethics statement (2026): optional, may sit in the body or the reference pages; recommended for sensitive data or tasks.
• Margins, font, and spacing are fixed by the template; editorial compression, not template tampering, is how you recover space.

Double-blind sweep

ECAI review is double-blind. Because AI papers lean on code, datasets, and named systems, the leak surface is wide:

# Mechanical pass on the submission PDF and any anonymized supplement
pdfinfo paper.pdf | grep -Ei 'author|creator|producer'
pdftotext paper.pdf - | grep -nEi 'github\.com/[a-z0-9-]+|gitlab|zenodo\.org/record|acknowledg|grant|funded by' | head
unzip -l supplement.zip | grep -Ei '\.git/|\.DS_Store|/home/|/Users/' | head
grep -rni 'university\|@[a-z0-9.]*\.edu\|our (system|group|lab)' supplement/ --include='*.md' | head

The AI-specific leaks: a method named after your group, a dataset or benchmark you are known for, an acknowledgements/funding line left in, a repository owner in a supplement, and self-citations phrased as "in our prior work [X]" instead of neutrally. Anonymize the supplement archive before upload.

Per-author cap and topics

• Submission cap: IJCAI-ECAI 2026 limits each author to no more than 8 submissions to the main track (verify per cycle). Count every paper on which a person is an author across your group.
• Topics/keywords: select the areas that match your evidence so the paper routes to the right area chair and reviewer pool — the single largest lever you control before reviews.
• Track choice: a deployed application paper may belong in PAIS (standalone editions) or an applied special track rather than the main track — decide before the abstract deadline (ecai-topic-selection).

Desk-risk triage

Finding at audit time	Severity	Real fix
Body over 7 pages	Desk-reject-grade	Cut or move to the supplement; reference pages do not absorb body text
Wrong template (acmart/IEEEtran/LNCS) or altered margins	Named desk-reject ground	Recompile in ijcai.sty/ecai.cls, recover space editorially
Identity leak in PDF, supplement, or method name	Anonymity violation	Re-anonymize and re-host; scrub PDF metadata
Abstract not registered by the earlier date	No submission slot exists	Nothing fixes this post-AoE — calendar it now
Author over the per-author submission cap	Submissions may be dropped	Prioritize and withdraw the weakest before the cutoff
Same work under review elsewhere	Dual-submission violation	Withdraw one venue; check the current concurrent-submission wording

Final-week order of operations

1. Freeze the 7-page body early; references can churn, the argument cannot.
2. Register title/abstract/authors/topics before the abstract cutoff.
3. Build the anonymized supplement (proofs, extra tables, code) and reference it from the body.
4. Run the mechanical double-blind checks on the final PDF and the final archive.
5. Confirm the template regime (ijcai.sty vs ecai.cls) and the exact page budget for THIS edition — do not assume 7+1 in a joint year.
6. Re-download the uploaded PDF and read it cold to confirm it is the file you meant.

Reverify each cycle

• Whether the edition is standalone (FAIA/ecai.cls, 7+1, PAIS) or joint with IJCAI (IJCAI proceedings/ijcai.sty, 7+2) — this flips template, publisher, and budget.
• All dates and AoE cutoffs; the per-author submission cap; the ethics-statement and generative-AI policy wording.
• Supplementary-material rules and whether an appendix is allowed.

Output format

[ECAI submission status] ready / blocked / needs work
[Regime] standalone ECAI (FAIA/ecai.cls) / joint IJCAI-ECAI (IJCAI/ijcai.sty)
[Abstract step] title/abstract/authors/topics locked by the earlier deadline? yes/no
[Format] pages used (body/refs), template compliance
[Anonymity] clean / leaks: <where>
[Caps & tracks] within per-author cap? correct track (main/PAIS)?
[Fix queue] <ordered, with owners and dates before the AoE cutoff>