Talk:Using Causal Graphs with Bayesian Optimization

The trade-offs and limitations section was a nice ending to the page! I also thought it was a good length!

Suggestions

• Maybe the background could be renamed as the problem or example since it sets up CBO and then maybe a separate background section could state what relevant knowledge is assumed or give some highlights of BO and causality that are important to understanding the papers
• Could also be clearer to state the second paper contributes POMIS and elaborate more on how it is an improvement
• Maybe POMIS could have its own section under Exploration Set (ES) for CBO
• May be nice to label the causal graph as Figure 1 so it can be (presented in Figure 1)

Some typos

• objective function function -> objective function
• Further -> Furthermore
• explorationSet: This set holds the elements of the. -> The sentence needs to be completed.

Thank you so much for the comments.

I have renamed the background section. I also added a section at the end of the article regarding the individual contributions of both papers and how one utilizes the other. I segregated POMIS and MIS as subsections. And have also corrected the typos mentioined.

The article was delightful to read and a well-suited topic for our class.

Overall, this is a wonderful page on Causal graphs with Bayesian Optimization. I learnt how it can be used for optimizing systems where there are complex causal relationships between them. I appreciate the description for each pseudo code, it will help anyone new to this topic. Here are a few suggestions/clarifications.

• What does sparsity of the graph mean?
• As a part of related pages, can you add some other pages on causality and bayesian optimization already specified on the course page.
• I was thinking that this can be applied to other methods. can you add some other examples at the end.
• I found the trade-offs that CBO balances very interesting but it seems to be a part of paper1. What is the main contribution of paper2, could you make that clearer?

Thank you so much for the comments.

• I have added the definition of the sparsity of a graph.
• I also added a section of Examples at the end of how this can be applied to performance optimization in Systems.
• I added another section at the end of the article, which states the exact contributions of both papers and how one utilizes the other.