First off, very nice work on continuing to host the DBLP exporter!
I found one small bug which you might be able to fix (or know who to contact?).
Some publications share the same value for foaf:homepage, seemingly if they are published in the same collection.
For example, the two papers:
Both have the same homepage:
...which is the DOI of the collection in which both are published. Since foaf:homepage is an inverse-functional property, these two papers (and all papers in the same collection) can be inferred to be the same.
An easy fix would be to weaken foaf:homepage to foaf:page or similar, but I guess that the DOI of most publications are probably unique. A more fine-grained option would be to run a script to detect and weaken only non-unique DOIs to foaf:page.