Skill UI

A comprehensive guide for researchers detailing how to establish rigorous provenance for on-device performance measurements (e.g., energy, latency, accuracy). It mandates recording all necessary context—including device model, SoC, OS build, framework versions, power instrument setups, and thermal conditions—to ensure results are verifiable and reproducible across different hardware or environments for academic publication.

Provides comprehensive guidelines for packaging and presenting research artifacts (code, data, models, benchmarks) submitted to NeurIPS. It details requirements for anonymous review packages, public release, and ensures scientific claims are reproducible by providing necessary metadata, licenses, and execution details.

This guide provides a comprehensive checklist for designing, executing, and presenting empirical evidence for academic papers in Pattern Recognition and AI. It covers critical areas like establishing fair baselines, ensuring statistical rigor (mean ± std dev, significance testing), conducting sufficient ablation studies, preventing data leakage, and reporting computational costs. It helps authors make their findings withstand rigorous external review.

A rigorous framework for designing reproducible and scientifically sound experiments in robotics and AI, specifically for academic publication (e.g., RSS). It guides users to shift focus from simple system demonstration to testing a falsifiable scientific claim. Key areas covered include deriving conditions from a claim, establishing hardware trial protocols, managing simulation vs. real-world evidence splits, and performing detailed failure attribution analysis.

This skill processes unstructured, rambling, stream-of-consciousness input (like dictation or brainstorming) into a highly structured 'echo' audit. It captures the user's core mission, locked decisions, open questions, and all model inferences, ensuring the user verifies the model's belief, not just the stated doubts. Ideal for refining fuzzy ideas before committing to work.

A specialized guide for authors aiming to submit to the Conference on Uncertainty in Artificial Intelligence (UAI). It provides expert advice on framing the core contribution—such as probabilistic reasoning, causal inference, and uncertainty modeling—and calibrating the manuscript for highly technical, statistical ML reviewers. Use this skill to refine your submission strategy, identify evidence gaps, and properly position your work relative to other top-tier AI venues like AISTATS and IJCAI.

A comprehensive guide for creating the two required artifacts for a WACV paper: an anonymous package for double-blind review, and a public release post-acceptance. It emphasizes reproducing deployed system claims (latency, power, robustness) and maintaining consistency across revision rounds.

A comprehensive guide for strengthening the reproducibility of academic papers, especially those involving systems claims (e.g., power usage, latency). It details maintaining a 'recipe ledger' for all reported metrics, ensuring constraint-aware measurement of environmental variables, and managing data/artifact consistency across the entire two-round revision process.

This tool benchmarks the recall of PII detection models, crucial for ensuring privacy compliance before deployment. It uses synthetic fixtures and enforces a recall floor by measuring label-aware exact-span and grapheme recall against predefined gold standards. It is essential for auditing models and preventing critical privacy leaks.

This skill evaluates de-identification and clinical NER models against rigorous leakage-first release gates (G1a-G8). Instead of relying on aggregate F1 scores, it focuses on detecting residual Protected Health Information (PHI) leakage and enforcing per-label recall floors, which is crucial for HIPAA compliance. Determines if a model is RELEASABLE or QUARANTINED.

This skill converts free-text radiology narratives (CT, MRI, X-ray, etc.) into structured, machine-readable data. It automatically identifies and segments key report sections (Technique, Findings, Impression) and extracts critical clinical details such as anatomy, measurements, laterality, BI-RADS/Lung-RADS scores, and follow-up recommendations. It is crucial for building structured medical records and decision support systems.

A comprehensive skill designed to transform contested questions or complex decisions into a structured, auditable evidence map. It systematically tracks all supporting, contradicting, qualifying, and missing evidence, ensuring that reasoning relationships and source regions are preserved for rigorous review.