Skill UI

A comprehensive guide for hardening deep learning papers, particularly in Computer Vision (CVPR). It details best practices for achieving rigorous reproducibility by utilizing the Compute Reporting Form (CRF), maintaining a detailed recipe ledger, ensuring benchmark split hygiene, and meticulously documenting all training hyperparameters, seeds, and variance analyses for academic publication.

A comprehensive guide for authors submitting to computer vision conferences, detailing how to structure, organize, and utilize supplementary material. It clearly distinguishes between core claims (which must be in the main body) and supportive evidence, such as full ablation grids, detailed derivations, and video results, which belong in the supplement. The guide covers best practices for handling anonymity, video evidence, and avoiding common submission bugs.

A comprehensive guide for academic researchers on how to structure and support claims in AI papers, particularly for top-tier conferences like ECAI. It details how to match the type of claim (e.g., completeness, efficiency, generalization) to the appropriate evidence—whether it requires a formal proof, a controlled empirical comparison, or a real-world deployment demonstration. The guide covers best practices for ML (seeds, ablations, baselines), multi-agent systems, and ensuring full provenance and academic honesty.

Comprehensive guidelines for structuring and submitting reproducible AI/ML research papers for top-tier conferences. It details two modes: Mode 1 for theoretical/KR work (requiring full, explicit proofs in the supplement), and Mode 2 for empirical/ML work (requiring fixed seeds, cached outputs, and rigorous provenance pinning for datasets, models, and environments). The guide also covers best practices for maintaining double-blind anonymity and properly dividing content between the main body and the supplementary material.

A comprehensive guide for authors submitting computer vision papers (e.g., to ECCV). It outlines best practices for ensuring the full reproducibility of AI research, covering everything from detailed training recipes and hyperparameter logging to pinning foundation model dependencies and reporting variance honesty. Essential for academic publication integrity.

A comprehensive guide for researchers submitting NLP papers, detailing best practices for achieving high reproducibility. It instructs users to pin model versions, log data provenance, report compute costs, and adhere rigorously to the Responsible NLP checklist. Essential for ensuring academic integrity and enabling rigorous verification of ML experiments.

This guide outlines best practices for structuring advanced NLP/AI research papers for academic submission. It details the critical differences between core content pages, mandatory limitations sections, discretionary appendices, and supplementary uploads, ensuring that reviewers are properly guided and that the work remains reproducible and verifiable.

A comprehensive guide for establishing academic rigor in AI and data science research. It outlines best practices for achieving reproducibility, emphasizing the mandatory release and documentation of raw data, trained models, and qualitative findings. Key tools include datasheets, model cards, and detailed provenance logging, ensuring that research claims are fully auditable and consistent.

This guide details the best practices for packaging, documenting, and releasing research artifacts (code, models, data) for top-tier ML conferences. It covers two stages: preparing a non-identifiable archive for anonymous review, and managing the public release runway, focusing on reproducibility, licensing, and sustained maintenance.

A comprehensive guide for structuring the "Related Work" section for top-tier computer vision conferences (ICCV, CVPR, ECCV). It covers best practices for citing intervening-cycle literature, managing concurrent preprints, ensuring double-blind anonymity, and correctly attributing source venues to maximize novelty perception and score highly with reviewers.

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.

This guide provides detailed best practices for ensuring the reproducibility of advanced AI research across various domains, including algorithms, theories, multi-agent systems, and datasets. It outlines the specific evidence required—from pseudocode and seeds to full proofs and public datasets—to elevate a submission's credibility rating from 'irreproducible' to 'convincing' for top-tier AI conferences.