Entities with no KG records

[golnazads]

The query that collects records to populate KG tables are limited to top planetary journals and limited year (recent), that might not have records for some of the entities (ie, Fimbulthul Catena). For the ones that the feature type is part of the feature name there should be no issue, since I think these entities are not ambiguous. So need to make sure these entities are identified and recognized as not having KG score.
@aaccomazzi

[aaccomazzi]

@sjarmak do you have any suggestions or insights?

[sjarmak]

@aaccomazzi @golnazads As a caveat, the full context of this is unclear to me and so I'd benefit from some clarification. Entities with the feature type as part of the name aren't necessarily ambiguous from just the name, right? As in, there's an "Adams crater" on Mars and on the Moon. Or, do you mean they are not ambiguous with additional contextual information that is separate from the KG score? I am unfamiliar with how the KG score is calculated, I couldn't identify the specifics associated with this score in Golnaz's paper. Is the plan to, for the cases where a KG score cannot be applied, proceed with the label and confidence based on the paper relevance and LLM chat scores only, perhaps with the confidence lowered by some amount to reflect that no KG score was provided? Is there any reason why the KG table needs to be populated in a way that restricts the number of planetary features that could be assigned a KG score?