Certainty-Validity: A Diagnostic Framework for Discrete Commitment Systems

The following includes the paper for the Certainty-Validity (CVS / CyVy) evaluation metric for datasets. CyVy helps with ML research by allowing to guide the model towards being confidentally correct and minimize confidentally inaccurate. In the repository are the following:

• "certainty_validity.py", a python script that allows for using the evaluation metric.
• "Certainty-Validity: A Diagnostic Framework for Discrete Commitment Systems", the main working paper for this.
• "cvs-supplementary-data.zip", a zip file that contains the a full 30 epoch benchmark in progress and epochs, with an additional figure of the isolated phase diagram that is at exists at the bottom of the paper.