Visualizing objective and acquisition function for current campaign

[BMGla5er]

I am interested in visualizing the acquisition function / recommender similar to how the mean and variance of the posterior can be (I made an example below for the three-hump camel function where the grey dashed line is the mean), however I am having trouble identifying the correct way to go about this. Has anyone else performed this visualization with Baybe before?

[Scienfitz]

Hello @BMGla5er

the interface to access the acqf values is not as straightforward as for the posterior statistics at the moment unfortunately, something nicer is already planned from our side

Until then, take a look at this example which i think is the closest published example how to access / visualize the acqf values. It uses a custom hook to extract the acqf values, however that might not even be needed for you if you just want a one time snapshot. If you work in a continuous space the only difference is that you need to generate a grid of values on which to evaluate the acqf first, instead of just using the candidates

[BMGla5er]

Thank you very much, @Scienfitz ! I'll take a look at the example.

[Scienfitz]

fyi @BMGla5er access to the target predictions and acquisition values is now even more convenient

[BMGla5er]

Thank you very much, @Scienfitz this looks very helpful and interesting to further explore!