KDD should be kicked out!

[lalalandlala]

Better late than never!

KDD is now a conference that top schools will not consider, and it is very misleading to rank it with other top machine learning conferences.

One solution is to use data mining instead of information retrieval, and let KDD replace SIGIR. Because information retrieval can no longer represent the most cutting-edge development direction of AI, and SIGIR lacks an audience even more than KDD.

[Sleepychord]

I think this issue is, to some extent, emotional and subjective.
In my opinion, KDD is still among the best conferences, but not the same audience as ICML et al.
1., I think the motivation of CSRanking is to "define" the good universities in CS research by objective metrics, instead of adjusting the scores based on the fame of the universities. The submission preference of a university / department depends on many factors, especially the topic distribution of the faculty.
2. KDD is now a conference that top schools will not consider is also not a fact. In CSRanking, Tsinghua (my university), Cornell and UIUC published the most KDD papers in past 5 years, and they are all very good universities even based on reputation. More important, KDD attracts many wonderful industrial submissions, which is slightly different from other venues. This unfortunately cannot be reflected in CSRanking.
3. but I am agree with the most parts of the solution. The audience of the KDD is more similar to WWW now. My suggestion is to replace "Web & IR" to "Data Mining & Information Retrieval", and move KDD into it, because the Web is slightly misleading for beginners.