How to reduce duplication of Works

[VladimirAlexiev]

This query about SemOpenAlex works:

PREFIX dct: <http://purl.org/dc/terms/>
PREFIX Service: <http://www.metaphacts.com/ontologies/platform/service/>
PREFIX entitylookup: <http://www.metaphacts.com/ontologies/platform/service/entitylookup/>
SELECT * WHERE {
  SERVICE Service:entityLookup {
    ?subject entitylookup:entityName "semopen";
             entitylookup:limit 100 ;
             entitylookup:score ?score;
             entitylookup:rank ?rank.
  }
  ?subject dct:title ?title
} ORDER BY DESC (?score) DESC (?rank) 

Returns 7 Works that are 3 "true" works, plus 4 variants/versions thereof:

subject	rank	title	comment
https://semopenalex.org/work/W4388144113	90.0	SemOpenAlex: The Scientific Landscape in 26 Billion RDF Triples	published
https://semopenalex.org/work/W4385682125	45.0	SemOpenAlex: The Scientific Landscape in 26 Billion RDF Triples	arxiv preprint
https://semopenalex.org/work/W4393797504	38.0	SemOpenAlex Embeddings	Zenodo "all versions"
https://semopenalex.org/work/W4393798967	38.0	SemOpenAlex Embeddings	Zenodo new version
https://semopenalex.org/work/W4393895619	38.0	SemOpenAlex Embeddings	Zenodo old version
https://semopenalex.org/work/W4393691335	36.0	RDF Knowledge Graph SemOpenAlex-SemanticWeb	Zenodo "all versions"
https://semopenalex.org/work/W4393746966	35.0	RDF Knowledge Graph SemOpenAlex-SemanticWeb	Zenodo only version

If this pattern holds, then over half of all Works in SOA are duplicates.
The problem is especially galling for the last case: every Zenodo resource, even if it has no versions at all, is represented by two Zenodo DOIs and consequently twice in SOA.

I know that these problems come from OpenAlex but I don't know where to report them.
I also understand that deduplicating works is not easy, eg the arxiv preprint doesn't state the DOI of the published version.
But maybe at least you can remove the Zenodo "all versions" URL/DOI, and maybe all old versions?