robertoemanuel posed the question "How many results will the UI receive?"

This is an unknown quantity at the present. The only limit we have is the maximum time the Counterfactual algorithm will execute; however in that time there is no limit on the number of intermediate results that could be returned.

Should we limit results to a maximum number (e.g. 15)?

A number of aspects need to be considered:

• Should we discard subsequent results returned from the explainer (even if the optimal solution is returned after the 15th)? The 15th result would be marked as the final.

• Should we stop the explainer once the 15th result is provided? There have previously been concerns about the feature to "stop" processing as it requires a broadcast mechanism to all nodes in a cluster/pod that could be handling the request. Systematically stopping the explainer has the same issues.