2 papers
cs.LG2026
DocDjinn: Controllable Synthetic Document Generation with VLMs and Handwriting Diffusion
Marcel Lamott, Saifullah Saifullah, Nauman Riaz +11
Effective document intelligence models rely on large amounts of annotated training data. However, procuring sufficient and high-quality data poses significant challenges due to the…
cs.CL2024
LAPDoc: Layout-Aware Prompting for Documents
Marcel Lamott, Yves-Noel Weweler, Adrian Ulges +3
Recent advances in training large language models (LLMs) using massive amounts of solely textual data lead to strong generalization across many domains and tasks, including documen…