sure, you can use these new tools to help you fill out existing forms of archival representation (named entities, for example), in ways that are often disappointing compared to human cataloging
-
-
Prikaži ovu nit
-
but the more interesting projects I think engage with, say, image vector spaces, as forms of archival representation themselves. Yes, those vectors need to be remediated into e.g. nearest neighbor search to be legible to humans
Prikaži ovu nit -
but those remediations actually provide a NOVEL representation - to be used in parallel with your finding aids, your hierarchies, your name authorities
Prikaži ovu nit -
this isn’t to say a human-computer loop for generating “traditional” archival representations like named entitites is useless - it can be really good (emphasis on the HUMAN in that loop, tho)
Prikaži ovu nit -
I just think it’s missing out on the fullest possibilities that computer-generated artefacts like vector spaces can offer if we engage with them as new representations in their own right, not something to be boiled down unsatisfactorily
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
I think I’ve generally opted for semi automated
-
there r more tweets in the threadddd PS been thinking much about your "conditions of possibility" lately... we gotta sneak you out to Pittsburgh at some point
- Još 1 odgovor
Novi razgovor -
-
-
I completely agree with this take. When I see some of the things we could do in the way of automated “metadata” extraction, _just_ filling in our existing metadata buckets seems too limited. It is interesting to start thinking about how we create new interfaces to this for users
-
Yeah! This is actually where I think technologists helping users and archivists better imagine "what do you WISH you could do" could help go beyond replicating existing representations (finding aid, EAD, "digital item") in re-skinned interfaces
- Još 1 odgovor
Novi razgovor -
-
-
Take a look at what the National Library of Norway is doing. They are doing deep ML dives into content to make the content, itself, searchable, rather than just the objects.https://www.nb.no/artikler/fantastic-futures/ …
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.
