ML Materials Predictor
Othersml-materials-predictorML Materials Predictor is an Agent Skill for general use published in the a5c-ai/babysitter repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Machine learning skill for nanomaterial property prediction and discovery acceleration
QUICK START
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/a5c-ai/babysitter/blob/HEAD/library/specializations/domains/science/nanotechnology/skills/ml-materials-predictor/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/ml-materials-predictor, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
ML Materials Predictor
Purpose
The ML Materials Predictor skill provides machine learning capabilities for accelerated nanomaterial discovery and property prediction, enabling data-driven approaches to materials design and optimization.
Capabilities
- Feature engineering for materials
- Property prediction models (GNN, transformers)
- Active learning for experiment design
- High-throughput virtual screening
- Synthesis success prediction
- Transfer learning for small datasets
Usage Guidelines
ML Materials Workflow
-
Data Preparation
- Collect and curate dataset
- Generate features (composition, structure)
- Handle missing values
-
Model Development
- Select appropriate architecture
- Train with cross-validation
- Evaluate on held-out test
-
Application
- Screen candidate materials
- Prioritize experiments
- Validate predictions
Process Integration
- Machine Learning Materials Discovery Pipeline
- Structure-Property Correlation Analysis
Input Schema
{
"dataset_file": "string",
"target_property": "string",
"model_type": "random_forest|gnn|cgcnn|megnet",
"features": "composition|structure|both",
"task": "train|predict|screen"
}
Output Schema
{
"model_performance": {
"mae": "number",
"rmse": "number",
"r2": "number"
},
"predictions": [{
"material": "string",
"predicted_value": "number",
"uncertainty": "number"
}],
"top_candidates": [{
"material": "string",
"predicted_property": "number",
"rank": "number"
}]
}