Skill UI

This guide outlines the rigorous methodology for ensuring that computational research results, especially those involving hardware architecture and simulation, are fully reproducible. It emphasizes pinning the entire result chain—from simulator commits and configuration files to workloads, sampling recipes, and statistics scripts—and mandates creating a manifest for every figure. Key practices include recording machine state for silicon runs and containerizing the environment to prevent data drift.

This comprehensive guide provides best practices for packaging supplemental materials and artifacts for top-tier academic conferences like ICDE. It instructs authors on how to build a robust, reproducible evidence chain, including parameterized workload generators, configuration scripts, and detailed logs. The goal is to ensure that reviewers can quickly and accurately validate all performance claims and technical contributions, making the work 'turnkey' and verifiable.

A comprehensive guide for ensuring the reproducibility of data mining research for top-tier conferences like ICDM. It addresses the unique challenges of triple-blind review and strict page limits by detailing necessary artifacts, such as configuration files, seed lists, and anonymized packages. Learn how to achieve Rerunnable, Rebuildable, or Attested reproducibility tiers to solidify your scientific claims.

A comprehensive guide for researchers submitting to major speech conferences (like INTERSPEECH). It details how to harden the reproducibility of speech results, covering crucial 'decay channels' such as corpus versions, text normalization rules (WER), objective metrics (EER), subjective protocols (MOS), and model stochasticity. The guide emphasizes pinning specific toolkits, commit hashes, configurations, and hardware details to ensure scientific rigor and verifiability.

A comprehensive guide for structuring the evidence bundle required for robotics research submissions. It details the critical components—including hardware ledgers, time-stamped logs, configuration files, code, and data—necessary to ensure that scientific claims are fully auditable and reproducible, addressing the needs of skeptical reviewers.

This guide provides best practices for ensuring the scientific rigor and reproducibility of data mining and machine learning research published at top-tier conferences like KDD. It details a three-tier honesty model (Rerunnable, Rebuildable, Attested) and provides comprehensive checklists for disclosing pipeline stages, handling stochasticity, and documenting configurations to build reader trust.

This guide details the process of preparing complex research artifacts (simulators, configurations, traces, and scripts) for post-acceptance evaluation, especially for microarchitecture papers. It outlines best practices for achieving ACM badges like 'Available' and 'Reproducible,' focusing on structuring the code base with tiered entry points (smoke, key, full) and handling dependencies, licensed data, and simulation time constraints.

This guide outlines advanced methodologies for achieving high reproducibility in networked-systems research. It details capturing crucial provenance ledgers—including topology, traffic traces, system configuration, and run details—to ensure that published results are verifiable and not dependent on ephemeral experimental conditions. It emphasizes characterizing inherent variance rather than demanding absolute determinism, providing a rigorous framework for academic publication.

A comprehensive guide and tool to upgrade legacy PowerShell Pester test suites across major versions (v3→v4, v4→v5, v5→v6). It covers structural changes, syntax updates, and breaking changes related to the Discovery/Run model, mock handling, and configuration updates. Use this when modernizing or fixing broken tests after Pester version bumps.

This specialized tool analyzes Git diffs against a stated development intent to detect scope creep. It identifies overly broad pull requests, unrelated files, dependency additions, public API renames, or configuration changes. It guides users on whether changes should be kept, split into separate commits, or if the connection requires explicit justification, ensuring clean and focused code merges.

Provides comprehensive guidelines for improving the reproducibility of academic network measurements and research artifacts, particularly for conferences like SIGCOMM. It details the necessary evidence map and the 'four ledgers'—topology, traffic, configuration, and run—to ensure that all reported claims can be accurately recreated and verified by the wider academic community.

A comprehensive guide for authors in data systems and related fields on achieving rigorous, verifiable reproducibility for academic papers. It mandates detailing the provenance of every claimed result by pinning five critical layers: code version, configuration parameters, dataset source, workload characteristics, and hardware specifications. Best practices include reporting distributions (e.g., percentiles, variance) and maintaining a detailed 'repro debt ledger' to withstand intense peer review.