Everything you need to know about materials optimization with surrogate models, active learning, and Pareto analysis.
# Install MatForge
$ pip install materia
# Initialize a water filtration project
$ materia init --domain water --name my-membrane
# Run optimization (500 evals, 15 rounds)
$ materia run material.yaml --budget 500 --rounds 15
# View Pareto-optimal results
$ materia results --top 10
Found 12 Pareto-optimal materials
Best PFOS rejection: 99.2%Create your first campaign in 5 minutes
Set up MatForge locally with Docker or pip
Materials, parameters, objectives, and Pareto fronts
YAML format for defining material optimization problems
Defining input variables with ranges and units
Setting optimization targets and directions
Hard and soft constraints on solutions
Covariance Matrix Adaptation Evolution Strategy
Neural network surrogate with MC Dropout uncertainty
Acquisition functions and convergence criteria
NSGA-II non-dominated sorting and crowding distance
11 materials science domains with physics models
Creating your own evaluator plugins
Domain-specific equations and parameters
Campaign management, results, and export endpoints
Real-time campaign progress updates
Programmatic access via the materia package
Scaffold a new material optimization project
Execute an optimization campaign
View and export campaign results
Generate an interactive HTML dashboard