brycewang-stanford ACM MM Workflow
Use this to run an ACM Multimedia submission as a scheduled campaign rather than a deadline scramble. The SIGMM calendar has a long gap between the April submission chain and the July decision, so the work clusters at the front and the back.
The ACM MM calendar (2026 anchors, AoE)
| Phase | 2026 date | What must be true | Skill |
|---|---|---|---|
| Scope + track choice | pre-March | Contribution is cross-modal; thematic area and track chosen | acmmm-topic-selection |
| Draft + evidence | March | First page leads with the multimedia claim; ablations planned | acmmm-writing-style, acmmm-experiments |
| Abstract | Mar 25 | Abstract registered in OpenReview | acmmm-submission |
| Paper | Apr 1 | 6–8 sigconf pages, anonymous, final PDF | acmmm-submission |
| Supplement | Apr 8 | Media/appendix package, anonymous, plays | acmmm-supplementary |
| Rebuttal | ~Jun 4 | Anonymous response, small confirmatory results | acmmm-author-response |
| Notification | ~early Jul | Decision + tier | acmmm-review-process |
| Camera-ready | ~Aug 6 | De-anonymized ACM version, artifact released | acmmm-camera-ready |
| Conference | Nov 10–14 | Registered; media-ready talk/poster | acmmm-camera-ready |
Treat these dates as the 2026 snapshot; reopen the current site each cycle. Note the 2026 notification date was listed inconsistently (July 7 on the CFP vs. July 9 elsewhere).
Front-loaded work (before April)
The April chain is unforgiving because three deadlines land in eight days. Do the slow work early:
- Lock the thematic area and track — they set format and blinding, and switching late is expensive.
- Build the cross-modal evidence: matched baselines and the leave-one-modality-out ablation that proves the fusion matters.
- Plan any user study now; recruiting raters cannot be done in the final week.
- Draft so the first page already leads with the multimedia contribution.
The three-deadline sprint
T-14d: freeze claims; media assets rendering from scripts
T-7d: body fits 6-8 sigconf pages; references on overflow pages only
Mar 25: abstract registered (do not miss the earliest deadline)
Apr 1: paper PDF final, anonymous, metadata-clean
Apr 8: supplement packaged, anonymous, media plays for a stranger
The long gap and the rebuttal
Between April and June, prepare the confirmatory experiments a reviewer is likely to ask for (the missing baseline, the extra ablation) so the June rebuttal is a matter of reporting results, not scrambling to run them.
Back end (July–November)
- On acceptance, de-anonymize, complete ACM rights/CCS, and release the public artifact.
- Register early and plan travel to Rio; visas and international logistics are the usual failure.
- Build a presentation that plays the media — the seam between modalities should be seen/heard, not just described.
- On rejection, use the reviews to re-target: MM Asia, ICMR, a workshop, or next year's MM.
Risk register
| Risk | Where it bites | Mitigation |
|---|---|---|
| Missing the abstract deadline | Mar 25 forecloses the paper | Register the abstract days early |
| Wrong track blinding | Desk reject | Confirm the track call before formatting |
| Fusion not shown to matter | Reviews / rebuttal | Run the leave-one-modality-out ablation pre-submission |
| User study not recruited in time | Perceptual claim unbacked | Start recruiting before April |
| Broken anonymous media link | Lost evidence in review | Test the mirror from a clean machine |
| Rio travel/visa delay | Cannot present | Start visa paperwork at acceptance |
Re-targeting after a rejection
An ACM MM rejection is not the end of the work. Match the reviews to the next move: a retrieval-leaning paper can go to ICMR; a systems-leaning one to MMSys; a regionally scoped or earlier-stage result to ACM Multimedia Asia or a workshop; a matured version to the ACM TOMM journal or next year's ACM MM. Fix the decisive critique — usually the cross-modal-significance one — before resubmitting anywhere.
Output format
[Current phase] <scope/draft/abstract/paper/supplement/rebuttal/decision/camera-ready/conference>
[On track?] yes / at-risk: <what slips>
[Next deadline] <date + deliverable>
[Active skill] <which skill to run now>
[Risks] <ordered>