Course talk:CPSC532:StaRAI2020:KnowledgeGraphs

Looks good.

The reason that I wanted the Ronaldo example or the Donald Trump one was that I wanted an example where we needed a reified entity to make it work. You say you need reification but none of the examples use it. (I think this is one of the problem with large scale KGs; that they try to avoid reification, often losing information. Perhaps include the dates that Ronaldo was captain of Portugal, or played for each club he played for. (See https://www.wikidata.org/wiki/Q11571)

Club-position is a weird property, as what if someone played different positions for different clubs?

On https://www.wikidata.org/wiki/Q11571 it is "member of sports team" (https://www.wikidata.org/wiki/Property:P54) that requires reification. I am not sure how it is actually represented in wikidata. (Think about what it is that ends on 1 July 2009.)

Thanks, If I want to include information about the time period then I think I will need reification.

Very clear and interesting article. I am not sure if the example of the Universality of Triples is a good starting point for the article. It may be helpful to add some lines about triples and their relationship with the knowledge graphs and then delve into the details and examples.

Thanks Melika, for your feedback. I will edit the article so that there is a smooth transition to triplets.

Great page! I really like the use of examples in building up the topic. My only advice would be to expand the section "Research and Tasks" to get a detailed image of the capabilities of knowledge graphs in terms of application, building, inferring, etc... I think the current version might be too concise.

Fair point. I did want to make it quite concise, and might have exaggerated. Figure 2 in this survey gives a good overview of that section -- and my writing was inspired by it.

Thanks for sharing the survey! The concise nature makes more sense now. This information might be hard to find so it's more of a bonus if that could be included:

I'm not sure if that survey or some other paper covers this, but including the current shortcomings/strengths in the four areas of research/tasks of knowledge graphs could be helpful.