Skill UI

This guide provides comprehensive best practices for designing and auditing empirical experiments for top-tier ML conferences. It covers necessary ablations, robustness checks (e.g., data shift, seeds), fair baselines, and detailed compute reporting to ensure that reported gains are scientifically genuine, reproducible, and address common reviewer skepticism.

A comprehensive self-auditing guide designed for authors preparing manuscripts for submission to top-tier academic journals. It systematically addresses recurring, structural, and critical flaws across theory contribution, econometric identification, mechanism explanation, heterogeneity analysis, and writing conventions. This tool helps authors proactively identify and correct potential weaknesses before facing peer review.

This guide provides a comprehensive framework for stress-testing Machine Learning experimental evidence before submission or rebuttal. It covers critical areas such as establishing strong baselines, conducting mechanism-isolating ablations, reporting variance, checking for data leakage, and disclosing compute costs. Use this to ensure your ML claims are scientifically sound, robust, and fair against rigorous academic scrutiny.

A comprehensive methodology guide for designing and auditing empirical studies in software engineering, specifically focused on maintenance, evolution, and change impact. It outlines best practices for ensuring evidence is proportional to the claim, managing data provenance (SHAs, criteria), handling statistical confounds (survivorship, effect sizes), and integrating rigorous LLM evaluation protocols.

A comprehensive guide for auditing and positioning the novelty and eligibility of research submissions in the field of software maintenance and evolution. It teaches authors to write a 'delta-first' related work section, contrasting their contribution against key prior works and covering specialized literature lanes (e.g., MSR, ICSE, TSE) to demonstrate deep command of the academic community.

This guide provides rigorous standards for designing, validating, and reporting security research, particularly for high-stakes academic conferences like IEEE S&P. It covers critical areas including adaptive adversary modeling, measurement validity (sampling, ethics), statistical reporting, establishing strong baselines/ablations, and adhering to ethical constraints (IRB, authorization) to ensure evidence is robust and reproducible.

This guide provides comprehensive standards for designing, auditing, and documenting AI experiments for major conferences like IJCAI. It covers establishing strong baselines, conducting ablation studies, and reporting meticulous details on hyperparameters, computational infrastructure, and ethical risks, ensuring maximum reproducibility and academic credibility for AI research.

A comprehensive checklist for auditing and preparing main-track conference submissions for top-tier AI venues like IJCAI. This guide covers critical compliance points, including double-blind anonymity enforcement, strict page limits, LLM usage policies, author details verification, and identifying potential desk-reject risks before final submission.

This guide offers detailed guidelines for assembling and auditing supplementary materials for top-tier AI conferences (e.g., IJCAI). It covers critical aspects like separating the technical appendix from resubmission files, ensuring anonymity across all artifacts, maintaining the self-contained nature of the core paper, and managing source code and data for maximum reproducibility.

This framework provides a comprehensive methodology for designing, auditing, and reporting empirical measurements about large-scale systems like the Internet. It guides users to ensure measurements are scientifically rigorous, by matching claims to evidence, quantifying coverage bias, designing for temporal changes, and adhering to strict ethical guidelines and data provenance standards.

A comprehensive guide for designing and auditing rigorous experiments for sensing systems and information processing claims. It emphasizes the necessity of measuring performance on real hardware (not simulation), accurately instrumenting ground truth, quantifying energy consumption, latency, and physical footprint, and establishing fair, skeptical baselines. This ensures scientific rigor and reproducibility in research submissions.

A comprehensive guide for designing, auditing, and reporting reliable robotic experiments, particularly for major conferences like IROS. It moves beyond simple success rates, detailing how to map every claim to verifiable evidence (trials, simulations, field data). Key areas covered include defining pre-set success criteria, reporting failure taxonomies, ensuring fair baselines, and quantifying the critical sim-to-real gap using advanced small-n statistical methods.