AutoRA Extrapolation Experimentalist
Project description
Extrapolation Experimentalist
The extrapolation sampling method identifies novel experimental conditions where the prediction of a model exhibits the highest slope compared to already existing data.
For each novel condition, denoted as $x_i$, with its corresponding prediction $y_{\text{pred}, i}$, the process begins by identifying the nearest existing datapoint, $x_{\text{nearest}, i}$, which has an associated observed value $y_{\text{existing}, i}$. The slope between these points is then calculated as follows:
$$ m_i = \frac{y_{\text{pred}, i}-y_{\text{existing}, i}}{x_i-x_{\text{nearest}, i}} $$
The condition with the highest slope is selected first:
$$ \underset{i}{argmax}(m_i) $$
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for autora_experimentalist_falsification-0.0.1.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | af6ff3a5ec410e8b4635a5f96f7f3aaf3d515998079007c9de0c9766caede488 |
|
MD5 | 69703ce4c8e1fca4383b23333fcaa40b |
|
BLAKE2b-256 | 810064b4e69dc57ef06ff89825395738f06c4cd7fdbeeb549ca2b4a61a45a848 |
Hashes for autora_experimentalist_falsification-0.0.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 817b9e10e733cb2db94616d7c461d309f407ea36e489dded3d328392fee0901b |
|
MD5 | 7b00daed57ecb86b3000606209c997f6 |
|
BLAKE2b-256 | 404cf2cebea5e72a68ca26d437cb5f29010e78cb6ae75cddadedee624c4298f0 |