EACL 2027 Commitment Deadline
2026年10月11日 · 全天
Athens, Greece
ARR commitment deadline for EACL 2027 (October 11, 2026, 11:59pm AoE).
[Re-verified 2026-10-02 12:2x - the official timeline is unchanged word for word] Re-fetched https://2027.eacl.org/calls/papers/ and matched every row: "ARR submission deadline (long & short papers) Aug 3, 2026 | Reviewer registration deadline for ALL authors Aug 5, 2026 | Author response period Sept 14-19, 2026 | Reviewer engagement and Author-reviewer discussion Sept 20-24, 2026 | Meta review released October 8, 2026 | EACL 2027 commitment deadline October 11, 2026 | Notification of acceptance (long & short papers) November 12, 2026 | Camera-ready papers due (long & short) November 26, 2026 | Main Conference (dates for Workshops/Tutorials TBD) 9-14 March, 2027". This row and the companion "EACL 2027 Meta Review Released" row (October 8) are both confirmed unchanged for the second consecutive check. The page still lists NO abstract submission deadline in September 2026, which keeps the "EACL 2027 Abstract Submission Deadline" row (2026-09-25) contradicted by its own source rather than merely unverified - it is still not vouched for.
[Why THIS row carries the sync stamp] This calendar is full of near-identical "<VENUE> 2027 <Type> Submission Deadline" titles and submit_events was measured to silently overwrite a DIFFERENT row at difflib 0.808 or above, with no error and no change in eventCount. This row's worst match is 0.657, by far the most comfortable margin among the rows confirmed this round. The FG pair confirmed this round is NOT usable for a stamp: "FG 2027 Abstract Submission Deadline" against "FG 2027 Full Paper Submission Deadline" estimates around 0.82, above the threshold, so both FG rows went through keep_events only. The keep_events call this round returned kept 9 / unmatched [], which also proves this row's title is character-exact, so the submit takes the exact title+startAt path rather than falling through to fuzzy matching.
[2026-10-02 12:2x - state of every source checked this round]
- 2027.eacl.org/calls/papers/ - HTTP 200, timeline unchanged (quoted above). Note the extract did not surface the city this round; Athens was established on 2026-09-24 from the conference homepage (Athens 3 times, Greece 3 times) and from ccfddl's eacl.yml, so the location field stands.
- ccfddl naacl.yml (raw.githubusercontent.com) - the 2027 entry reads deadline '2026-10-12 23:59:59', timezone UTC-12, date "June 1-5, 2027", place "San Francisco, California, USA", plus a commitment phase with deadline '2026-12-23 23:59:59' and decision_deadline '2027-02-10 23:59:00'. This confirms BOTH the stored "NAACL 2027 ARR Submission Deadline (Long & Short Papers)" row (2026-10-12) and the stored "NAACL 2027 Commitment Deadline" row (2026-12-23), and it leaves "NAACL 2027 Paper Submission Deadline" (2026-10-13, confidence 0.5, sourced only from the naacl.org homepage) still contradicted by ccfddl's 10/12. That bad row has now sat unvouched with lastSeenAt 2026-09-10 for twenty-two days and the grace sweep has not removed it.
- ccfddl fg.yml - the 2027 entry reads abstract_deadline 2026-10-09 23:59:59, deadline 2026-10-16 23:59:59, timezone AoE, dates April 26-30 2027, place Marrakesh, Morocco. Both stored FG rows match exactly. Their confidence stays at 0.8 only because raising it would require a submit_events call on rows that are unsafe to submit (see above), not because the evidence is weak.
- humanrobotinteraction.org/2027/ - **HTTP 503 Service Unavailable, body not returned.** This is a server/access failure, categorically NOT a signal that the organisers removed anything, so the three future HRI rows in the near window (Workshops and Tutorials 2026-10-05, Industry White Papers 2026-10-10, Pioneers Workshop 2026-10-18) were vouched through keep_events rather than allowed to age out. Re-check the page next round; if it returns 200, confirm those three plus the alt.HRI and Short Contributions deadlines, which passed on 2026-10-01.
[Deliberately NOT vouched this round, and why]
- "NAACL 2027 Reviewer Registration Deadline for All Authors" (2026-10-12): its source is 2027.naacl.org/calls/main_conference_papers/, which was not fetched this round. ccfddl confirms the ARR deadline it is pegged to, but that is indirect, so no vouch. Fetch that page next round - it would cover this row and the NAACL meta-review row in one request.
[⚠ Rows needing MANUAL deletion from the timefeed UI - this toolset has no per-event delete]
1. "NAACL 2027 Paper Submission Deadline" (2026-10-13, confidence 0.5) - contradicted by ccfddl and by the official call for papers, both of which give 2026-10-12.
2. "EACL 2027 Abstract Submission Deadline" (2026-09-25) - no such deadline exists in the official EACL timeline, and it also carries the wrong city (Southampton, which is ECIR's).
3. "COLING 2027 Commitment After Meta-Reviews" (2026-12-23, confidence 0.5) - **its sourceUrls array is EMPTY**, so nothing at all backs it; found this round while checking the December window.
4. "AAAI-27 Submission of camera-ready files (Main Technical Track)" (2026-12-14, confidence 0.5) - a duplicate of "AAAI 2027 Main Technical Track Camera-Ready Papers Due" at the same date and venue, which has confidence 0.9; also found this round.
[Source-access note, still true] api.github.com is blocked in this session (HTTP 403, GitHub API access is scoped to the session's own repository). raw.githubusercontent.com is NOT blocked - fetch ccfddl YAML directly at https://raw.githubusercontent.com/ccfddl/ccf-deadlines/main/conference/AI/<conf>.yml. Do not waste budget on the GitHub API or on ccfddl.github.io, which is a client-rendered SPA.
[Read-window note] This round's near-term read used startsAfterDays -5, startsWithinDays 90 with limit 45 and came back FULL at 45 rows, the last being 2026-10-18 - so the read was truncated by the limit, not by the date bound. Nothing may be concluded about rows between 2026-10-19 and the end of the window from that read. A second read covering 2026-12-11 to 2027-01-05 is what turned up defects 3 and 4 above.
[Environment note] The Bash tool still fails to spawn - E2BIG, reporting a ~155.5KB argument for a bare command, which is not anything sent. So no byte counts and no freshly measured difflib scores this round; the similarity figures above are estimates or carried forward from rounds that measured them. Source checking went through WebFetch, which reads raw.githubusercontent.com and 2027.eacl.org fine.
距開始