Question: are there corpora size limits?

[JulianKingman]

I noticed in the paper says "...our benchmark, MINE, currently measures performance on relatively short corpora, whereas KGs are primarily used to handle massive amounts of information efficiently", are there limits to the size of the corpora, or is it just the benchmark that's limited?

Practically speaking, could this be run on a collection of hundreds of books worth of information, or would that be difficult? Would the knowledge graph get confused and over-fit?

[CQUPT-CZL]

I noticed in the paper says "...our benchmark, MINE, currently measures performance on relatively short corpora, whereas KGs are primarily used to handle massive amounts of information efficiently", are there limits to the size of the corpora, or is it just the benchmark that's limited?我注意到论文中写着“......我们的基准 MINE 目前衡量相对较短的语料库的性能，而 KG 主要用于有效处理大量信息“，语料库的大小是否有限制，或者只是基准有限？

Practically speaking, could this be run on a collection of hundreds of books worth of information, or would that be difficult? Would the knowledge graph get confused and over-fit?实际上，这是否可以在数百本书的信息集合上运行，还是会很困难？知识图谱会变得混乱和过度拟合吗？

Yes, I think due to the reasons of time complexity and cost, as well as the context length limitations of large models