We're always looking for new ways to discover and access our collection. The recent development of large language models has created new opportunities by encoding human language into a format that's easy for computers to work with. One of those formats is embeddings.
Embeddings is a way to represent words as a list of numbers.
It means we can encode a human readable word like 'hello' into something a computer can quickly compare to lots of other words, finding matches.
This collection explorer uses embeddings to show you algorithmically matching works as you tap and explore.
See what the machine sees / see what the algorithm recommends...
Read more about it at: labs.acmi.net.au