Consider including AISTATS in AI field for broader topic coverage

[fsaad]

Dear Maintainers,

At present, the AI field contains only two venues (AAAI; IJCAI) and ML contains three (ICML; ICLR; NeurIPS). A noteworthy omission for AI is the International Conference on Artificial Intelligence and Statistics (AISTATS), which is a top long-running venue for diverse topics including statistical methods, Bayesian techniques, approximate inference, causal learning, optimization, nonparametrics, numerical methods, AI theory, etc. The need for a more representative theory venue, specifically mentioning AISTATS, has also been noted in #5428 and #5677.

Admittedly, I do not know the formal inclusion criteria for csrankings. At the surface level:

• AISTATS submissions have grown by ~500% in the last 10 years; matching ICML and outpacing IJCAI.
• AISTATS 2024 acceptance rate is 27.6%, similar to several current venues; e.g., ICLR 2024 is 30.5% and ICML 2024 is 27.5%.
• Google Scholar lists AISTATS among the top venues in terms of h5-index and h5-median statistics, keeping in mind AISTATS has much fewer accepted total papers as compared to the other conferences in that list.

I do appreciate the need to minimize the number of changes to this list to maintain stability. However, it seems worth including a more representative top AI venue that covers important topics which are not traditionally captured by IJCAI/AAAI (on the AI side) or by NeurIPS/ICML/ICLR (on the ML side).