Skill UI

Conversational research ideation assistant that helps scientists generate novel hypotheses, explore interdisciplinary analogies, challenge assumptions, and outline experiment pathways during early-stage planning.

Processes machine learning experiment outputs to build comparison tables, compute statistics, and surface insights with recommendations for follow-up trials whenever users ask to analyze results or make comparisons.

This skill automates the generation of publication-quality figures and comparison tables for academic papers. It converts structured experimental data (e.g., JSON, CSV) into various plots, including line graphs, bar charts, scatter plots, and heatmaps, ensuring consistent styling suitable for LaTeX submissions. It handles the core data visualization needs for ML and research reports, greatly streamlining the process of creating professional results sections.

This skill provides a multi-round, high-depth critical review of academic research projects, simulating the rigorous feedback process of top-tier conferences (e.g., NeurIPS, ICML). It guides the user through identifying logical gaps, suggesting missing experiments, and refining the core narrative claims to maximize acceptance chances.

Guides on using SwanLab for open-source ML experiment tracking, offering local or self-hosted dashboards, rich media logging, and integrations with PyTorch, Transformers, Lightning, or Fastai for comparing runs and monitoring metrics.

Trackio is a comprehensive library for logging, visualizing, and managing ML experiment metrics and alerts. It allows developers to use a Python API to log metrics and set diagnostic alerts (INFO/WARN/ERROR) during training. Furthermore, a dedicated CLI enables metric retrieval and dashboard viewing, syncing data to Hugging Face Spaces for persistent, remote monitoring, making it ideal for autonomous ML agent workflows.

This gate systematically evaluates experimental outcomes against intended research claims. It collects metrics from various sources (W&B, logs, etc.) and uses an objective judgment model to determine if the data supports, partially supports, or fails to support the hypothesis. It provides specific evidence gaps and suggests concrete next steps, guiding the research workflow before writing papers or finalizing designs.

This expert tool guides users through rigorous statistical analysis, specializing in A/B testing, hypothesis testing, and interpreting experimental results. It helps define sample sizes, executes appropriate tests (Z-test, T-test, Chi-square), and critically distinguishes between 'statistical significance' and 'practical significance' to ensure data-driven decision-making for product and business strategy.

This skill provides comprehensive commands to manage the entire AI model evaluation lifecycle using Arize. Users can list, retrieve metadata, and export results from experiments. It is ideal for comparing multiple model runs, benchmarking performance against a dataset, and analyzing model accuracy using structured evaluation workflows (CRUD operations).

This tool functions as a persistent, structured knowledge base for the entire research lifecycle. It accumulates and maps all core research entities—including published papers, generated ideas, experiments, and testable claims—along with their complex relationships (e.g., contradiction, support, inspiration). Inspired by LLM Wiki patterns, it ensures that knowledge is compounded and perpetually queryable, moving beyond single-use literature reviews.

A powerful tool designed for academic researchers and scientists. It searches a curated database of scientific literature and retrieves structured metadata and quantitative experimental data (including methods, results, sample sizes, and quality scores) directly from full-text papers, going far beyond simple abstracts. Ideal for systematic literature reviews, evidence synthesis, and meta-analyses requiring detailed, comparable study information.

Guides structured hypothesis generation by combining literature searches, evidence synthesis, competing explanations, and experimental design, with mandatory AI-generated scientific schematics for complete reports.