Welcome to CoFI!

Welcome to CoFI!#

Welcome to the CoFI Documentation! CoFI, the Common Framework for Inference, is an open-source tool that bridges the gap between domain expertise and inference, within and beyond the field of geoscience.

Whether you’re a student, an educator, or an industry professional, CoFI is your go-to tool. It simplifies the process of applying inference techniques, allowing for a wide range of applications from academic research and teaching to solving real-world geoscience problems in the industry. With its user-friendly, flexible and accessible interface, CoFI empowers you to focus on what truly matters - the science.

        graph TD;
    cofi(CoFI - Common Framework for Inference):::cls_cofi;
    parameter_estimation(Parameter estimation):::cls_parameter_estimation;
    linear(Matrix based solvers):::cls_parameter_estimation;
    non_linear(Optimization):::cls_parameter_estimation;
    linear_system_solvers(Linear system solvers):::cls_parameter_estimation;
    linear_solverlist(scipy.linalg.lstsq <br>...):::cls_solvers;
    optimization(Non linear):::cls_parameter_estimation;
    optimization2(Linear):::cls_parameter_estimation;
    opt_solverlist(scipy.optimize.minimize <br> torch.optim <br> ROL <br>...):::cls_solvers;
    ensemble_methods(Ensemble methods):::cls_ensemble_methods;
    direct_search(Direct Search):::cls_ensemble_methods;
    amc(Monte Carlo):::cls_ensemble_methods;
    amc_solverlist(Neighbourhood Algorithm <br> Bayesian Optimization <br> Slime mold algorithm <br> Border Collie Optimization <br> ...):::cls_solvers;
    ng(Deterministic):::cls_ensemble_methods;
    ng_solverlist(Nested grids <br> Hilbert Curves<br>...):::cls_solvers;
    bs(Bayesian Sampling):::cls_ensemble_methods;
    mcmc(McMC samplers):::cls_ensemble_methods;
    mcmc_solverlist(Basic metropolis<br>Affine Invariance sampler<br>emcee <br> pyMC <br> ...):::cls_solvers;
    rjmcmc(Trans-D McMC):::cls_ensemble_methods;
    rjmcmc_solverlist(Basic Trans-D <br> BayesBay <br> RJ-mcmc):::cls_solvers;

    cofi --> parameter_estimation;
    parameter_estimation --> linear;
    linear --> linear_system_solvers;
    linear_system_solvers -.- linear_solverlist;
    parameter_estimation --> non_linear;
    non_linear --> optimization;
    non_linear --> optimization2;
    optimization -.- opt_solverlist;
    optimization2 -.- opt_solverlist;

    cofi --> ensemble_methods;
    ensemble_methods --> direct_search;
    direct_search --> amc;
    amc -.- amc_solverlist;
    direct_search --> ng;
    ng -.- ng_solverlist;
    ensemble_methods --> bs;  
    bs --> mcmc;
    mcmc -.- mcmc_solverlist;
    bs --> rjmcmc;
    rjmcmc -.- rjmcmc_solverlist;

classDef cls_cofi fill: #d4a373, stroke-width:0;
classDef cls_parameter_estimation fill: #ccd5ae, stroke-width:0;
classDef cls_ensemble_methods fill: #e9edc9, stroke-width:0;
classDef cls_solvers fill: #faedcd, stroke-width:0;

# CoFI API has flexible ways of defining an inversion problem. For instance:
import cofi

inv_problem = cofi.BaseProblem()
inv_problem.set_objective(my_objective_func)
inv_problem.set_initial_model(my_starting_point)

# Once a problem is defined, `cofi` can tell you what inference tools you can
# use based on what level of information you've provided:
inv_problem.suggest_tools()   # a tree will be printed

# Run an inversion with these lines:
inv_options = cofi.InversionOptions()
inv_options.set_tool("torch.optim")
inv_options.set_params(options={"num_iterations": 50, "algorithm": "Adam"})

inv = cofi.Inversion(inv_problem, inv_options)
result = inv.run()
print(result.success)
print(result.model)