1 paper · 1 filter
Bram Willemsen, Gabriel Skantze
In this paper, we explore the use of a text-only, autoregressive language modeling approach for the extraction of referring expressions from visually grounded dialogue. More specif…