AeSlides: Incentivizing Aesthetic Layout in LLM-Based Slide Generation via Verifiable Rewards DOI: XXXXXXX.XXXXXXX Conference: arXiv; Under Review; April 2026 ISBN: 978-1-4503-XXXX-X/2018/06 529 CCS: Computing methodologies Natural language processing CCS: Computing methodologies Reinforcement learning CCS: Applied computing Multi / mixed media creation CCS: Human-centered computing Visualization design and evaluation methods CCS: Computing methodologies Computer vision
Yiming Pan 1 Chengwei Hu 2 Xuancheng Huang 2 Can Huang 2 Mingming Zhao 2 Yuean Bi 2 Xiaohan Zhang 2 Aohan Zeng 2 Linmei Hu 1 · 2026
Read AeSlides: Incentivizing Aesthetic Layout in LLM-Based Slide Generation via Verifiable Rewards DOI: XXXXXXX.XXXXXXX Conference: arXiv; Under Review; April 2026 ISBN: 978-1-4503-XXXX-X/2018/06 529 CCS: Computing methodologies Natural language processing CCS: Computing methodologies Reinforcement learning CCS: Applied computing Multi / mixed media creation CCS: Human-centered computing Visualization design and evaluation methods CCS: Computing methodologies Computer vision by Yiming Pan 1 Chengwei Hu 2 Xuancheng Huang 2 Can Huang 2 Mingming Zhao 2 Yuean Bi 2 Xiaohan Zhang 2 Aohan Zeng 2 Linmei Hu 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-22840.txt.
Description
Pan and eight colleagues present AeSlides, a reinforcement learning framework that rewards language models directly for aesthetic slide layouts instead of relying on costly visual reflection or weak fine-tuning signals. Verifiable metrics score layout problems cheaply, and a GRPO method optimizes generation against them. Trained with only 5,000 prompts, it raises aspect-ratio compliance from 36 percent to 85 percent while cutting whitespace, collisions, and imbalance, and human judges rate its slides better overall.
Catalog record
| CREATOR | Yiming Pan 1 Chengwei Hu 2 Xuancheng Huang 2 Can Huang 2 Mingming Zhao 2 Yuean Bi 2 Xiaohan Zhang 2 Aohan Zeng 2 Linmei Hu 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84002 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.22840