SlidesGen-Bench: Evaluating Slides Generation via Computational and Quantitative Metrics

2026-01-14

AI Summary

Existing evaluation methods for slide generation systems lack cross-architectural comparability and often rely on subjective or uncalibrated judgments. To address this gap, this work proposes the first benchmark that simultaneously ensures generality, quantifiability, and reliability by treating generated outputs as visual renderings and enabling end-to-end quantitative assessment across three dimensions: content, aesthetics, and editability. The approach is agnostic to the underlying generation model and integrates computational visual analysis with multidimensional metrics. Furthermore, it introduces Slides-Align1.5k, a dataset aligned with human preferences. Experiments across nine representative systems and seven distinct scenarios demonstrate that the proposed benchmark achieves significantly higher correlation with human judgments than existing evaluation protocols.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Evaluation and Analysis
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success

Abstract

The rapid evolution of Large Language Models (LLMs) has fostered diverse paradigms for automated slide generation, ranging from code-driven layouts to image-centric synthesis. However, evaluating these heterogeneous systems remains challenging, as existing protocols often struggle to provide comparable scores across architectures or rely on uncalibrated judgments. In this paper, we introduce SlidesGen-Bench, a benchmark designed to evaluate slide generation through a lens of three core principles: universality, quantification, and reliability. First, to establish a unified evaluation framework, we ground our analysis in the visual domain, treating terminal outputs as renderings to remain agnostic to the underlying generation method. Second, we propose a computational approach that quantitatively assesses slides across three distinct dimensions - Content, Aesthetics, and Editability - offering reproducible metrics where prior works relied on subjective or reference-dependent proxies. Finally, to ensure high correlation with human preference, we construct the Slides-Align1.5k dataset, a human preference aligned dataset covering slides from nine mainstream generation systems across seven scenarios. Our experiments demonstrate that SlidesGen-Bench achieves a higher degree of alignment with human judgment than existing evaluation pipelines. Our code and data are available at https://github.com/YunqiaoYang/SlidesGen-Bench.

Problem

• slide generation
• evaluation benchmark
• quantitative metrics
• LLM evaluation
• human preference alignment

Innovation

• SlidesGen-Bench
• quantitative evaluation
• slide generation
• human preference alignment
• computational metrics