Thanks very much for your research. In the second paper, you mentioned that restricting the sample size is needed (partly) because of the computation issue. I am curious about if we still use the very large 40 million sample size, are there any methods to make the computation tractable? Thanks!
Thanks very much for your research. In the second paper, you mentioned that restricting the sample size is needed (partly) because of the computation issue. I am curious about if we still use the very large 40 million sample size, are there any methods to make the computation tractable? Thanks!