TY  - RPRT
TI  - Text, Pixels, or Both? Evaluating Input Representations for Multimodal Document QA
AU  - Nikhil Reddy Pottanigari
AU  - Sepideh Kharaghani
AU  - Saverio Vadacchino
AU  - Alejandro Posada
AU  - Kurt MacDonald
AU  - Ying Zhang
PY  - 2026
UR  - https://arxiv.org/abs/2609.22628
ID  - 2609.22628
ER  -