Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2604.11969 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914470086836224 |
|---|---|
| author | Ozden, Tarik Can VS, Sachidanand Horoz, Furkan Kara, Ozgur Kim, Junho Rehg, James Matthew |
| author_facet | Ozden, Tarik Can VS, Sachidanand Horoz, Furkan Kara, Ozgur Kim, Junho Rehg, James Matthew |
| contents | We introduce ArcDeck, a multi-agent framework that formulates paper-to-slide generation as a structured narrative reconstruction task. Unlike existing methods that directly summarize raw text into slides, ArcDeck explicitly models the source paper's logical flow. It first parses the input to construct a discourse tree and establish a global commitment document, ensuring the high-level intent is preserved. These structural priors then guide an iterative multi-agent refinement process, where specialized agents iteratively critique and revise the presentation outline before rendering the final visual layouts and designs. To evaluate our approach, we also introduce ArcBench, a newly curated benchmark of academic paper-slide pairs. Experimental results demonstrate that explicit discourse modeling, combined with role-specific agent coordination, significantly improves the narrative flow and logical coherence of the generated presentations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_11969 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Narrative-Driven Paper-to-Slide Generation via ArcDeck Ozden, Tarik Can VS, Sachidanand Horoz, Furkan Kara, Ozgur Kim, Junho Rehg, James Matthew Artificial Intelligence We introduce ArcDeck, a multi-agent framework that formulates paper-to-slide generation as a structured narrative reconstruction task. Unlike existing methods that directly summarize raw text into slides, ArcDeck explicitly models the source paper's logical flow. It first parses the input to construct a discourse tree and establish a global commitment document, ensuring the high-level intent is preserved. These structural priors then guide an iterative multi-agent refinement process, where specialized agents iteratively critique and revise the presentation outline before rendering the final visual layouts and designs. To evaluate our approach, we also introduce ArcBench, a newly curated benchmark of academic paper-slide pairs. Experimental results demonstrate that explicit discourse modeling, combined with role-specific agent coordination, significantly improves the narrative flow and logical coherence of the generated presentations. |
| title | Narrative-Driven Paper-to-Slide Generation via ArcDeck |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2604.11969 |