Tag
This paper presents a training-free graph-based framework for reading order inference in complex document layouts, using language model signals and a max-regret inference rule. The method significantly outperforms existing baselines on historical manuscripts and multi-column benchmarks, achieving 95% edge accuracy on wrap-around Glossa layouts.
ArcDeck is a multi-agent framework that generates presentation slides from academic papers by modeling logical flow through discourse trees and iterative agent refinement, outperforming direct summarization methods. The paper introduces ArcBench, a new benchmark for evaluating paper-to-slide generation with emphasis on narrative coherence and logical structure.