2 papers
cs.LG2025
Zero-Shot Document Understanding using Pseudo Table of Contents-Guided Retrieval-Augmented Generation
Hyeon Seong Jeong, Sangwoo Jo, Byeong Hyun Yoon +3
Understanding complex multimodal documents remains challenging due to their structural inconsistencies and limited training data availability. We introduce \textit{DocsRay}, a trai…
cs.CL2025
Not All Personas Are Worth It: Culture-Reflective Persona Data Augmentation
Ji-Eun Han, Yoonseok Heo
Incorporating personas into conversational AI models is crucial for achieving authentic and engaging interactions. However, the cultural diversity and adaptability of existing pers…