Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/cikm-review-process
Use when reasoning about CIKM peer review — the EasyChair double-blind pipeline, the mixed IR/data-mining/knowledge-management reviewer pool, per-track evaluation criteria, the ACM Peer Review Policy including the no-AI-written-reviews rule, notification timing, and what actually moves borderline decisions.
Passed 15 of 22 test cases with the skill, and 3 of 22 without it.
and 13 more in the full report
14 fixed1 passed either way5 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 14 test cases the skill fixed. Same model, same prompt, with and without the skill.
Our CIKM submission received uniform middling scores across all three reviewers, with everyone agreeing the paper is clear but the contribution is thin. We plan to fix this by rewriting the introduction and abstract for a resubmission. Diagnose this outcome in standard review format.
| Checked | Without skill | With skill |
|---|---|---|
| uses the tag [Sharpest objection] | Missed | Passed |
| identifies uniform middling scores as an underlying evidence problem | Missed | Passed |
| advises strengthening core experimental evidence before resubmitting | Passed | Passed |
The instructions your model reads when the skill loads.
CIKM review runs on EasyChair under double-blind rules, inside the ACM Peer Review Policy, and — its defining feature — in front of a reviewer pool drawn from three communities at once. Verified 2026 mechanics (source map, 2026-07-08): submissions closed in May/June, notification lands August 7, and referees are explicitly barred from using AI systems to write reviews. Whether 2026 includes an author response window is unconfirmed in either direction (待核实); plan without assuming one.
A CIKM paper is typically read by people whose default standards differ:
| Reviewer's home lane | What they instinctively grade | Complaint they file most |
|---|---|---|
| Information retrieval | Evaluation design, baselines, metric discipline | "Baselines are stale / significance untested" |
| Data mining | Mechanism novelty, scalability, ablation logic | "Delta over the nearest KDD-line method unclear" |
| KM / databases | Data model, integration cost, system realism | "Would not survive real schema/scale/noise" |
The practical consequence: a paper optimized for one lane can draw its harshest
review from another. Write the submission so each lane finds its own checklist
satisfied — a defensible evaluation, an isolated mechanism, and a credible data
story — or explicitly scope the claim to the lanes it serves. This is also why the
reviewer-pool breadth rewards two-lane papers (see cikm-topic-selection): they
give more of the panel a reason to champion.
The five tracks are judged against different success definitions. Full Research is graded on novelty plus evidence depth; Short Research on the sharpness of a single finding, not breadth; Applied Research on the credibility of deployment evidence (launch, data release) and transferable lessons; Resource on documentation, licensing, and likely reuse; Demonstration on what a visitor can actually do with the prototype. Reviewing a submission against the wrong track's bar — the most common self-review error — produces both false confidence and false alarm.
test cases passed, vs 3 / 22 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.