Skill UI

This skill calculates model interpretability for algorithmic trading signals using PageRank. It constructs a feature-contribution graph from an LSTM or Transformer signal and executes a single-entry PageRank calculation. The result identifies the top-K features that most significantly drove the prediction, providing a "regulator-grade" audit trail essential for compliance with regulations like the EU AI Act.

This guide assists authors targeting the ACM SIGMOD conference. It provides expert advice on positioning, framing, and structuring database systems papers. It emphasizes strong algorithmic or systems contributions, realistic workloads, and necessary evidence, while also advising on routing relative to similar venues like VLDB and ICDE.

A comprehensive guide for academics submitting AI and statistical research papers. It details how to map every claim (theoretical, algorithmic, or empirical) to verifiable evidence, covering necessary details like datasets, hyperparameters, random seeds, compute resources, and uncertainty estimates. It outlines the standards for achieving high levels of reproducibility, from descriptive to turnkey, ensuring scientific rigor for top-tier conferences.

A comprehensive framework for analyzing and presenting algorithmic contributions in academic papers. It guides users beyond mere computational speed, requiring rigorous proof of complexity (Big O), convergence (rate and conditions), or approximation bounds (rho). Essential for optimizing and validating complex computational models in scientific fields.

This guide outlines the rigorous analytical methodology required for Operations Research (OR) manuscripts. It instructs users on selecting the appropriate mathematical machinery—such as duality, dynamic programming, convergence analysis, or concentration inequalities—to establish provable guarantees, approximation factors, and algorithmic complexity. It ensures that claims are backed by rigorous proofs, moving beyond mere heuristics.

This skill guides authors on strategically positioning data mining research for the SIAM International Conference on Data Mining (SDM). It helps adjust manuscript framing, emphasizing mathematical rigor, statistical depth, and algorithmic contribution suitable for top Computer Science venues. Use it for venue selection, understanding reviewer expectations, and structuring evidence to meet the high standards of SIAM-style research.

A comprehensive guide for designing and auditing visualization experiments for high-stakes academic submissions (e.g., IEEE VIS). It details how to match evidence—such as controlled studies, power analyses, and benchmarks—to the specific type of claim (perceptual, algorithmic, system). Learn best practices for statistical rigor, CVD-safe encoding, and maintaining full provenance to ensure your findings are scientifically sound and trustworthy.