Skip to main content

The goal of this project is to examine the performance of a mixture experimentalist–a hybrid of different experimental sampling strategies–in terms of its ability to recover a ground-truth model from synthetic data.

Project description

AutoRA Mixture Experimentalist

The Mixture Experimentalist identifies novel experimental conditions under which a hybrid of different experimental sampling strategies is used. This mixture can include any custom strategies such as falsification, novelty, crucial experimentation, uncertainty, elimination, aesthetic preferences, and arbitrary preferred/dispreferred regions of the space. The selection of conditions is based on a weighted sum of the scores obtained from these strategies.

Quickstart Guide

You will need:

Mixture Experimentalist is a part of the AutoRA package:

pip install -U autora["mixture_experimentalist"]

Check your installation by running:

python -c "from autora.experimentalist.sampler.mixture_experimentalist import mixture_sampler"

Usage

The Mixture Experimentalist can be used to select experimental conditions based on a mixture of different strategies. Here's a basic example:

from autora.experimentalist.sampler.mixture import mixture_sampler

# Define your condition pool, temperature, samplers, and parameters
condition_pool = ...
temperature = ...
samplers = ...
params = ...

# Use the mixture_sampler to select conditions
selected_conditions = mixture_sampler(
    condition_pool=condition_pool,
    temperature=temperature,
    samplers=samplers,
    params=params,
    num_samples=10
)

In this example, condition_pool is the pool of experimental conditions to evaluate, temperature controls the randomness of the selection (close to 0 for deterministic, higher for more random), samplers is a list of sampler functions with their weights in the mixture, and params is a dictionary of parameters for the sampler functions.

For more detailed usage instructions and examples, please refer to the documentation: https://github.com/blinodelka/mixture_experimental_strategies/blob/main/docs/basic-usage.ipynb.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mixture_experimentalist-1.0.0a1.tar.gz (155.3 kB view hashes)

Uploaded Source

Built Distribution

mixture_experimentalist-1.0.0a1-py3-none-any.whl (4.7 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page