Review for NeurIPS paper: Computing Valid p-value for Optimal Changepoint by Selective Inference using Dynamic Programming

Summary and Contributions: Update after reading the author rebuttal: I'd like to thank the authors for their detailed rebuttal. I appreciate that they aim to expand the experimental section and related work, and simplify the notation and descriptions where possible. I certainly think these changes can make the paper even more accessible and hope that the authors will also incorporate the other suggestions made to improve the manuscript. With regards to the runtime shown in Figure 7, I am still not convinced this is "almost linear": connecting the mean of the boxplots does not seem to give a linear relationship, but more likely a quadratic one. I'd recommend the authors fit a polynomial to this data and work out whether it is linear or quadratic in the number of samples.