CONIKS enables automated trust establishment with untrusted communication service providers by having the service provider maintain an auditable directory of all of its users' keys.

It's a public key directory scheme based on transparency logs. Part of me wonders whether a scheme like this could be adapted into a DID method, but it looks like this one (at least) requires a certain amount of structure in the user IDs, to unambiguously associate each ID with a particular "service provider" (their term).