Few-shot adaptation with 50 target-domain samples enables VLMs finetuned on different DocVQA domains to match or surpass fully supervised counterparts, revealing visual understanding as the primary bottleneck.
In: IEEE ICCST (2026)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Comparative Study of Domain-adapted VLMs for General Document Visual Question Answering
Few-shot adaptation with 50 target-domain samples enables VLMs finetuned on different DocVQA domains to match or surpass fully supervised counterparts, revealing visual understanding as the primary bottleneck.