kdense-ml-ai-selected
精选 K-Dense Scientific Agent Skills 的机器学习与人工智能工具包。用于 scikit-learn、PyTorch Lightning、Transformers、SHAP、time series ML、TimesFM、PyTorch Geometric、UMAP 等科研建模路线,并按 Auto-sci-research 的体育地理、城市暴露、空间公平、健康影响和论文解
SKILL.mdskills/kdense-ml-ai-selected/SKILL.md
GitHubK-Dense ML/AI Selected
This wrapper packages selected Machine Learning & AI skills from K-Dense-AI/scientific-agent-skills for Auto-sci-research.
Use it when a task needs technical playbooks for:
scikit-learnpytorch-lightningtransformersshaptimesfm-forecastingtorch-geometricumap-learn
Local Adaptation
Use these upstream skills with local research constraints:
- Define prediction target, population, spatial unit, temporal unit, and leakage risks before model choice.
- Prefer simple baselines before complex models.
- Separate predictive utility from causal interpretation.
- For exposure, accessibility, and spatial equity work, document spatial and temporal validation splits.
- Treat SHAP, feature importance, and embeddings as interpretation aids, not causal evidence.
For domain-specific guidance, also read:
../agent-auto-sci-ai-ml/references/k_dense_ml_ai_mapping.md../agent-auto-sci-ai-ml/references/ml_leakage_and_validation.md
Must Not Do
- Do not report only accuracy without calibration, uncertainty, and validation design.
- Do not mix training and test geographies or time periods without disclosure.
- Do not present model explanations as mechanisms unless the design supports that claim.