RecSys Submission
Run this audit before anything is uploaded to the RecSys submission system. RecSys is a
single-community but multi-track venue: the same manuscript can be legal in one track and
desk-rejectable in another, so the audit begins with "which track is this PDF for?" and only
then checks format. Reopen the current edition's Call for Contributions — every number below
is a 2026-cycle anchor (verified 2026-07-09), not a permanent rule.
Pick the track before the format
For 2026 RecSys reorganized its lineup: it dropped the Late-Breaking Results (LBR) poster
track and added Research & Practice (R&P) Notes. Route the contribution first.
| Track |
2026 budget |
What belongs here |
| Main - long paper |
8 content pages |
A rounded recommendation contribution with offline (and ideally online) evidence |
| Main - short paper |
4 content pages |
One focused, complete finding |
| Main - Past, Present & Future |
4 content pages |
Reflective/forward-looking position on a recommendation topic |
| Reproducibility |
see current CFP |
Repeating, refuting, or re-scoping prior results; datasets/frameworks for future reproduction |
| Industry |
see current CFP |
Deployed systems, production constraints, live A/B evidence |
| Resource / Dataset |
see current CFP |
A community dataset or software resource with build methodology |
| R&P Notes (posters) |
see current CFP |
Short findings/experiences; replaces LBR for 2026 |
References are excluded from the page count; appendices count inside it. Use the ACM
two-column template unmodified — no margin, font, or spacing surgery.
Anonymization for recommender papers specifically
Generic double-blind advice misses the leaks recommender papers actually carry:
- Proprietary logs and platform. "Interaction logs from our video platform in region X"
can name the company. Keep the statistics, genericize the source.
- Deployed system names. If the ranker already runs in production under a public name, the
name de-anonymizes; rename for review or argue from the current policy on preprints.
- The anonymous repository. RecSys expects a code/data link inside the paper rather than
a supplement upload — mirror it to an anonymized host and strip the owner, commit authors, and
cluster paths from history.
- Self-citations in third person; acknowledgements and funding removed; PDF metadata
stripped.
Blocking risks
- Wrong track for the contribution (a deployed-system paper squeezed into a Main long paper, or
a reproduction study submitted as a novel-method paper).
- Over budget once appendices are counted inside the limit.
- Identity leak via logs, platform, deployed system name, or repository owner.
- Untuned baselines or a random split on sequential data — not a format desk reject, but the
fastest route to a low score (see
recsys-experiments).
- Parallel submission of the same work to another venue.
Desk-reject and triage table
| Trigger |
Severity at RecSys |
Repair window |
| Over budget (appendix counted in) |
Desk reject |
None after the deadline |
| Template tampering |
Desk reject or chair flag |
None |
| Identity leak in PDF, logs, or repository |
Desk reject |
None |
| Wrong track's submission form |
Desk reject or forced move |
Usually none |
| Parallel/dual submission |
Desk reject at any time |
None |
| Baselines untuned; random split on session data |
Review-stage damage |
Only before the deadline |
Final-week sequence for an offline-evaluation paper
- Freeze the split: regenerate every ranking table from the versioned dataset and a temporal
split so the PDF and the anonymous repository agree.
- Re-run baseline tuning last under an equal budget; a stale, under-tuned baseline is the
classic reviewer catch when the numbers move.
- Compress into the content-page budget by demoting per-dataset analyses to the appendix or
repository, never by shrinking floats.
- Anonymity sweep: PDF metadata, acknowledgements, log provenance, deployed-system name,
repository owner.
- Confirm the submission-form abstract matches the PDF abstract and the track is correct.
# Pre-upload checks that catch the silent killers
pdfinfo paper.pdf | grep -iE "author|creator" # metadata identity leak
grep -inE "acknowledg|our (platform|service|company)" *.tex # textual / platform leaks
pdftotext paper.pdf - | grep -icE "tuned|grid search|equal budget" # baseline tuning stated?
pdftotext paper.pdf - | grep -icE "temporal split|leave-one-last" # leakage-aware split?
Format anchors
- RecSys uses the ACM two-column proceedings template; dense ranking tables and multi-panel
offline/online figures overflow columns fast, so compress early rather than on deadline night.
- The page counts above describe the 2026 cycle; treat every number as provisional and recheck
the current Call for Contributions before relying on it.
Output format
[RecSys readiness] Ready / Needs fixes / Wrong track / Not ready
[Track] main-long / main-short / past-present-future / reproducibility / industry / resource / R&P-note
[Budget check] <pages used>/<budget>, appendix-inside-budget confirmed yes-no
[Anonymity risks] <logs / platform / deployed name / repository owner / metadata>
[Evidence risk] <baseline tuning / split leakage / offline-only claim>
[Fix order] <ordered blocking fixes>
Source: brycewang-stanford/Awesome-Journal-Skills → RecSys-Skills/skills/recsys-submission/SKILL.md
1---2name: recsys-submission3description: Use when auditing an ACM RecSys submission before the deadline — routing between the Main, Reproducibility, Industry and Resource tracks, the ACM two-column content-page budgets with appendices counted inside, double-blind anonymization for recommender papers, the anonymous-repository expectation, and desk-reject triage.4---567# RecSys Submission89Run this audit before anything is uploaded to the RecSys submission system. RecSys is a10single-community but multi-track venue: the same manuscript can be legal in one track and11desk-rejectable in another, so the audit begins with "which track is this PDF for?" and only12then checks format. Reopen the current edition's Call for Contributions — every number below13is a 2026-cycle anchor (verified 2026-07-09), not a permanent rule.1415## Pick the track before the format1617For 2026 RecSys reorganized its lineup: it **dropped the Late-Breaking Results (LBR)** poster18track and added **Research & Practice (R&P) Notes**. Route the contribution first.1920| Track | 2026 budget | What belongs here |21|---|---|---|22| Main - long paper | 8 content pages | A rounded recommendation contribution with offline (and ideally online) evidence |23| Main - short paper | 4 content pages | One focused, complete finding |24| Main - Past, Present & Future | 4 content pages | Reflective/forward-looking position on a recommendation topic |25| Reproducibility | see current CFP | Repeating, refuting, or re-scoping prior results; datasets/frameworks for future reproduction |26| Industry | see current CFP | Deployed systems, production constraints, live A/B evidence |27| Resource / Dataset | see current CFP | A community dataset or software resource with build methodology |28| R&P Notes (posters) | see current CFP | Short findings/experiences; replaces LBR for 2026 |2930References are **excluded** from the page count; **appendices count inside** it. Use the ACM31two-column template unmodified — no margin, font, or spacing surgery.3233## Anonymization for recommender papers specifically3435Generic double-blind advice misses the leaks recommender papers actually carry:3637- **Proprietary logs and platform.** "Interaction logs from our video platform in region X"38 can name the company. Keep the statistics, genericize the source.39- **Deployed system names.** If the ranker already runs in production under a public name, the40 name de-anonymizes; rename for review or argue from the current policy on preprints.41- **The anonymous repository.** RecSys expects a code/data link **inside the paper** rather than42 a supplement upload — mirror it to an anonymized host and strip the owner, commit authors, and43 cluster paths from history.44- **Self-citations** in third person; **acknowledgements and funding** removed; **PDF metadata**45 stripped.4647## Blocking risks4849- Wrong track for the contribution (a deployed-system paper squeezed into a Main long paper, or50 a reproduction study submitted as a novel-method paper).51- Over budget once appendices are counted inside the limit.52- Identity leak via logs, platform, deployed system name, or repository owner.53- Untuned baselines or a random split on sequential data — not a format desk reject, but the54 fastest route to a low score (see `recsys-experiments`).55- Parallel submission of the same work to another venue.5657## Desk-reject and triage table5859| Trigger | Severity at RecSys | Repair window |60|---|---|---|61| Over budget (appendix counted in) | Desk reject | None after the deadline |62| Template tampering | Desk reject or chair flag | None |63| Identity leak in PDF, logs, or repository | Desk reject | None |64| Wrong track's submission form | Desk reject or forced move | Usually none |65| Parallel/dual submission | Desk reject at any time | None |66| Baselines untuned; random split on session data | Review-stage damage | Only before the deadline |6768## Final-week sequence for an offline-evaluation paper69701. Freeze the split: regenerate every ranking table from the versioned dataset and a **temporal**71 split so the PDF and the anonymous repository agree.722. Re-run baseline tuning last under an **equal budget**; a stale, under-tuned baseline is the73 classic reviewer catch when the numbers move.743. Compress into the content-page budget by demoting per-dataset analyses to the appendix or75 repository, never by shrinking floats.764. Anonymity sweep: PDF metadata, acknowledgements, log provenance, deployed-system name,77 repository owner.785. Confirm the submission-form abstract matches the PDF abstract and the track is correct.7980```bash81# Pre-upload checks that catch the silent killers82pdfinfo paper.pdf | grep -iE "author|creator" # metadata identity leak83grep -inE "acknowledg|our (platform|service|company)" *.tex # textual / platform leaks84pdftotext paper.pdf - | grep -icE "tuned|grid search|equal budget" # baseline tuning stated?85pdftotext paper.pdf - | grep -icE "temporal split|leave-one-last" # leakage-aware split?86```8788## Format anchors8990- RecSys uses the ACM **two-column** proceedings template; dense ranking tables and multi-panel91 offline/online figures overflow columns fast, so compress early rather than on deadline night.92- The page counts above describe the 2026 cycle; treat every number as provisional and recheck93 the current Call for Contributions before relying on it.9495## Output format9697```text98[RecSys readiness] Ready / Needs fixes / Wrong track / Not ready99[Track] main-long / main-short / past-present-future / reproducibility / industry / resource / R&P-note100[Budget check] <pages used>/<budget>, appendix-inside-budget confirmed yes-no101[Anonymity risks] <logs / platform / deployed name / repository owner / metadata>102[Evidence risk] <baseline tuning / split leakage / offline-only claim>103[Fix order] <ordered blocking fixes>104```105106---107108**Source:** [`brycewang-stanford/Awesome-Journal-Skills`](https://github.com/brycewang-stanford/Awesome-Journal-Skills) → `RecSys-Skills/skills/recsys-submission/SKILL.md`
thedixitjain