The record, exactly as recorded
Single-venue deadline pages, like OpenCurious's ICML 2027 page, live or die on precision, so here is our venue record with nothing added.
| Recorded field | ICML in our directory |
|---|---|
| Venue | International Conference on Machine Learning (ICML) |
| Tier | A* (CORE) |
| Tentative deadline month | January (confidence: high) |
| Event month | July |
| Host city | Seoul, South Korea |
| Acceptance rate | Not recorded (only 8 of 329 venues in the directory carry one) |
What tentative means
Our directory records a deadline month and a confidence level, not a day. January at high confidence means the cycle is stable enough to plan a submission calendar around, and not precise enough to book travel against. The exact day, the abstract-versus-full-paper split and the supplementary deadline exist only in the official call; check the conference site or the OpenCurious page above before treating any January date as real. What the record does settle is the shape of the year: a January deadline for a July conference gives roughly six months from submission to presentation.
ICML among the anchors
ICML opens the anchor year. The same records put NeurIPS deadlines in May for a December event in Sydney, ICLR in September for April in Rio de Janeiro, and CVPR in November (confidence: medium, and our record carries no host city for it) for June. An ML author can therefore run two anchor cycles a year without the deadlines colliding, which is exactly the planning fact the four-anchor chart shows. Our Submission Odds Checker holds the month and tier for all 141 conferences in the directory, so the same check works for the non-anchor venues too.