Skill UI

A comprehensive guide for researchers preparing academic artifacts for voluntary, post-acceptance evaluation (e.g., HPCA). It outlines best practices for structuring submissions for cold-start evaluators, defining reproducibility tiers (full reproduction, regeneration from logs, inspection), and managing computational budgets to successfully earn IEEE reproducibility badges.

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 skill provides comprehensive guidance for submitting research to npj Computational Materials. It defines the journal's scope, requiring rigorous computational methods (DFT, MD, ML, high-throughput screening) that yield generalizable materials insights or predictive capabilities. It educates authors on the required standards, including methodological rigor, reproducibility, and adherence to FAIR data principles, helping authors re-frame their work for a high-impact venue.

A comprehensive guide for ensuring robust scientific reproducibility in robotics research (RSS). It structures evidence into distinct 'Replication Tiers' (e.g., Rerunnable, Auditable), distinguishing between the computational half (code, seeds) and the physical half (robot, environment, human protocol). It mandates the use of platform ledgers and raw trial logs to convert claims of trust into verifiable, auditable evidence.