Skill UI

Builds AutoML pipelines by analyzing user requests, generating code for data preparation, model selection, hyperparameter tuning, validation, and metrics, then documenting artifacts for streamlined automated machine learning workflows.

Collection of battle-tested PyTorch training recipes covering architecture selection, optimizers, LR scheduling, precision tips, and debugging to train or fine-tune LLMs, vision, diffusion, and biomedical models with high GPU throughput.

Build, evaluate, and audit right-censored or competing-risk survival workflows with scikit-survival 0.28.0, covering leakage-safe pipelines, model selection, time-dependent prediction, and censoring-aware metrics such as IPCW concordance, dynamic AUC, and Brier scores.

Provides guided statistical analysis for research data, covering test selection, assumption checking, effect sizes, and APA reporting. Supports frequentist and Bayesian methods including t-tests, ANOVA, regression, and correlation. Ideal for comparing groups, testing hypotheses, and analyzing experimental or survey data to produce robust, reviewer-ready results.

This skill provides comprehensive methods and code examples for optimizing spending on Bright Data services. It covers product selection (Proxy vs. API), implementing caching strategies, bulk usage monitoring, and setting budget alerts, ensuring cost efficiency in large-scale web scraping projects.

Query and download public cancer imaging data (CT, MR, PET, pathology) from the NCI Imaging Data Commons. Supports metadata SQL lookups, DICOM access via idc-index, REST or MCP, cohort selection, browser visualization and license checks without authentication.

A comprehensive guide covering the canonical rules for Power BI report design. It enforces consistency across design identity, visual hierarchy, color discipline, chart selection, and accessibility. Focuses on structuring reports to answer specific questions and minimizing cognitive load, ensuring data effectiveness over mere aesthetics.

A comprehensive guide for strengthening the methodological design of advanced geographical manuscripts. It covers various paradigms including spatial/quantitative GIScience, remote sensing, qualitative human geography, and mixed methods. Key advice areas include rigorously addressing spatial dependence (MAUP), validating measurements, justifying case selection, and ensuring holistic integration of diverse data sources.

A comprehensive checklist for auditing academic submissions to ACM Multimedia (ACM MM). Covers thematic area selection, formatting constraints (6-8 page limit), enforcing double-blind anonymity, managing OpenReview mechanics, and identifying common desk-reject triggers. Essential for authors preparing technical papers for top-tier conferences.

This guide provides comprehensive econometric methodologies for establishing causality in accounting and finance research. It covers advanced techniques such as Difference-in-Differences (DID), Propensity Score Matching (PSM), Heckman selection models, and Instrumental Variables (IV), emphasizing the correct handling of regulatory changes and external shocks to meet top-tier academic standards.

This guide provides a comprehensive, advanced strategy for academic researchers writing microeconometric papers for top-tier journals (e.g., AEJ: Applied). It teaches authors how to anticipate sophisticated, technical objections from expert referees (such as issues with causal inference design, selection bias, or parallel trends). The focus is on pre-empting these objections by integrating solutions directly into the manuscript text, rather than waiting for a formal revision letter.

This guide outlines how to structure a rigorous robustness program for economic policy manuscripts. It emphasizes defending the overall policy takeaway, rather than merely reporting coefficient significance. Users learn to address common econometric threats—such as specification changes, pre-trends, multiple testing, and selection on unobservables—using advanced techniques (e.g., Wild-cluster bootstrap, Oster bounds) to ensure the policy conclusion is stable and credible.