This requires issue 24 to be complete. Tuning BKBs to produce some ground truth distribution (perhaps one found in the literature) means altering the conditional probabilities of the most sensitive factors in the BKB to the desired targets, whilst minimizing the overall change of the model. Once BKBs are tuned we can evaluate our model by reasoning over other known distributions as a validation process.
This requires issue 24 to be complete. Tuning BKBs to produce some ground truth distribution (perhaps one found in the literature) means altering the conditional probabilities of the most sensitive factors in the BKB to the desired targets, whilst minimizing the overall change of the model. Once BKBs are tuned we can evaluate our model by reasoning over other known distributions as a validation process.
See https://github.com/di2ag/Connections-Hypothesis-Provider-Roadmap/milestone/6 for other open issues.