Skip to content

Reproduce the mabuc_paradox figure from a library helper #4

@raphaelrrcoelho

Description

@raphaelrrcoelho

The launch figure (causal ~0.75 vs naive ~0.50 on identical interventional means) is generated by a throwaway script. Add a small documented helper or a notebook cell under examples/ that produces the running-average-reward curves from CausalThompsonSampling / NaiveThompsonSampling, so the result is reproducible from the library.

Files: examples/.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions