kdd-experiments
Use when designing or auditing the empirical section of a KDD paper, where evidence combines quality deltas with scalability and efficiency measurements, temporal-leakage-safe splits, mechanism-isolating ablations, tuning-symmetric baselines, and, for the ADS track, post-launch measurement design that survives the desk check.
Profile →npx skills add brycewang-stanford/Awesome-Journal-Skills --skill kdd-experiments --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# KDD Experiments Use this while the experimental plan is still cheap to change. A KDD empirical section answers four questions in order: does the method win, **at what scale**, at what computational price, and **because of which mechanism**? Papers that answer only the first question read as ML-flagship rejects retargeted at KDD — a pattern this venue's reviewers name openly. ## The four-axis evidence plan | Axis | Minimum credible evidence | Upgrade that wins reviews | |---|---|---| | Quality | Headline metric vs tuned baselines on named datasets with stated sizes | Multiple data regimes (sparse/dense, small/large, static/drifting) showing where the method does and does not help | | Scale | Largest-dataset run with hardware stated | Scaling curve (time and memory vs data size) with the complexity claim overlaid | | Efficiency | Wall-clock and memory vs baselines, same hardware | Throughput per component, so the O(·) claim is checkable per stage | | Mechanism | One ablation removing the claimed contribution | Full mechanism matrix: each named design decision toggled independently | A missing axis should be a stated limitation, never a silent hole. ## Data hygiene the practitioner-
- The four-axis evidence plan
- Data hygiene the practitioner-reviewers hunt for
- Baseline discipline
- Ablation logging that isolates mechanisms
- ADS-track measurement design
- Vignette: auditing a recommendation paper's evidence
- Reporting floor
- Output format
What does the kdd-experiments skill do?
Use when designing or auditing the empirical section of a KDD paper, where evidence combines quality deltas with scalability and efficiency measurements, temporal-leakage-safe splits, mechanism-isolating ablations, tuning-symmetric baselines, and, for the ADS track, post-launch measurement design that survives the desk check.
How do I install it?
Run `npx skills add brycewang-stanford/Awesome-Journal-Skills --skill kdd-experiments --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From brycewang-stanford/Awesome-Journal-Skills, a repository with 909 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.