Pith. sign in

REVIEW 3 major objections 5 minor 7 cited by

PPTAgent: Generating and Evaluating Presentations Beyond Text-to-Slides

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Edit-based agent beats text-to-slides on all three judged quality axes.

desk verdict A genuinely new edit-based presentation system with a solid dataset and evaluator, but the headline 'significantly outperforms' claim is not yet supported because the main judge shares criteria with the method and significance tests are missing. read the letter →

arxiv 2501.03936 v3 pith:BFDN4W6J submitted 2025-01-07 cs.AI cs.CL

Hao Zheng , Xinyan Guan , Hao Kong , Jia Zheng , Weixiang Zhou , Hongyu Lin , Yaojie Lu , Ben He
show 2 more authors
Xianpei Han Le Sun
This is my paper · ORCID
classification cs.AIcs.CL
keywords presentationgenerationslideeditingLLMagentMLLMjudgeevaluationdocument-to-slidesself-correctionreference-guided
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

PPTAgent argues that presentation generation should be treated as a reference-guided editing task rather than a text-to-slides summarization task. Given an input document and an example deck, it clusters the deck's slides by function, extracts a content schema for each cluster, drafts an outline, and then issues executable commands that rewrite the reference slides one element at a time. The paper also introduces PPTEval, a three-dimension judge (Content, Design, Coherence) using a multimodal LLM, and reports that PPTAgent outperforms rule-based and template-based baselines on all three. This matters because prior methods optimize text in isolation and produce visually flat decks, whereas an edit-based pipeline preserves a designer's layout while replacing content. If accepted, the work reframes the task from generation to curation and offers a reference-free way to measure deck quality.

What carries the argument

The load-bearing mechanism is the edit-API plus HTML-rendering loop. Instead of generating slides from scratch, the LLM reads a slide rendered as simplified HTML, selects a reference slide by its functional cluster, and writes small executable actions—replace_span, replace_image, clone_paragraph, del_span, del_image—that preserve the original layout and style. A REPL executes the actions, returns errors, and lets the model correct itself until the slide is valid. The other central object is PPTEval, which decomposes quality into slide-level Content and Design and presentation-level Coherence, with detailed scoring prompts; it supplies the numeric evidence for the paper's claims.

What would settle it

A direct human preference test in which raters compare unlabeled decks from PPTAgent, DocPres, and KCTV, plus a coherence-only study where opening and ending slides are removed or shuffled, would settle whether the reported gains are real or an artifact of the judge rewarding structural cues.

Watch

Extended reading notes

Core claim

At the paper's center is a workflow that turns slide creation into code. Stage I clusters reference slides into structural and content types and extracts a per-cluster schema of text and image elements; Stage II generates an outline that pairs each new slide with a reference slide and a document section, then emits sequences of API calls such as replace_span and replace_image against an HTML-rendered version of the slide, with a REPL supplying execution feedback for self-correction. PPTEval scores the resulting decks on Content, Design, and Coherence on a 1-to-5 scale, using GPT-4o as judge after converting slides to images. The headline empirical finding is that the Qwen2.5 + Qwen2-VL configuration reaches a 95% success rate and an average score of 3.67, beating DocPres and KCTV on all three judged dimensions, with the largest reported gap in coherence (4.48 versus 3.57 and 3.28).

Load-bearing premise

The load-bearing premise is that PPTEval's automated-judge scores are a valid proxy for human quality; the paper's own data show average human correlation of 0.71 but only 0.55 for coherence, and no direct human side-by-side comparison between PPTAgent and the baselines is reported.

Editorial extensions

If this is right

  • A deck generated this way inherits the visual layout of a human-designed reference slide, so the model does not need to specify positions, colors, or sizes by hand.
  • When the reference analysis includes structural slides such as openings, closings, and section headers, coherence scores jump; removing the outline drops coherence from 4.48 to about 3.36, making structural placement the main driver of judged coherence.
  • Replacing the HTML-rendered representation with the earlier code-interaction format lowers the success rate from 95.0% to 74.6%, so simplifying the verbose XML is what makes reliable editing possible.
  • The paper's weak correlations of PPL, ROUGE-L, and FID with judged content and design indicate that existing automatic metrics cannot substitute for a purpose-built presentation judge.
  • Generated quality is bounded by the quality of the reference deck, which means reference selection is itself a quality lever and a potential failure point.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because human-judge correlation is weakest for coherence (Pearson 0.55), the paper's largest claimed gain is also its least certain; a direct human test focusing on storyline flow, rather than the presence of opening and closing slides, would clarify whether the coherence advantage is substantive.
  • The cluster-extract-edit cycle is general: the same schema-and-edit loop could be applied to document-to-poster, document-to-infographic, or document-to-storyboard generation, where preserving template design matters.
  • Since quality is bounded by the reference deck, retrieving the best-matching reference per input document may matter as much as the generation model; a natural extension is retrieval-based reference selection instead of a fixed reference set.
  • PPTEval's absolute scores on human-authored decks are not reported, so a calibration check is to run the judge on the original Zenodo10K presentations; if human decks score below generated ones, the metric's scale needs recalibration.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes PPTAgent, a two-stage, edit-based system for generating presentations from documents: Stage I analyzes reference presentations by clustering slides into structural and content types and extracting content schemas; Stage II generates an outline and applies executable edit actions to selected reference slides, with a REPL-based self-correction mechanism. The paper also introduces PPTEval, an MLLM-as-a-judge evaluation framework scoring presentations on Content, Design, and Coherence, and releases Zenodo10K, a new dataset of over 10,000 presentations. Experiments compare PPTAgent with DocPres and KCTV across 500 generation tasks, reporting that PPTAgent significantly outperforms both baselines on all three PPTEval dimensions, together with ablations, a score-distribution analysis, self-correction statistics, and a human-LLM correlation study.

Significance. If the empirical claims hold, the paper makes a useful step beyond text-to-slides by formulating presentation generation as reference-guided editing, and it provides valuable public assets in the form of code, a dataset, and an evaluation framework. The ablations, multiple model configurations, and the human correlation study are careful and concrete. However, the headline comparison rests on an author-defined judge whose coherence rubric explicitly rewards structural slides and background information, which are exactly the components PPTAgent adds; the human correlation for coherence is only 0.55, the weakest dimension. The significance of the central claim therefore depends on ruling out an evaluation artifact, which the current evidence does not do.

major comments (3)
  1. [§3, Appendix F (Figs. 20, 23)] The headline coherence gains in Table 3 (4.48 vs. 3.24–3.57) are confounded with the method. The PPTEval coherence rubric awards 4 points for "basic background information (e.g., speaker or acknowledgments/conclusion)" and 5 points for "detailed and comprehensive background information," and the extraction prompt in Fig. 20 explicitly pulls speaker and date from opening and closing slides. PPTAgent's Stage I/II is designed to detect and reproduce structural slides (Opening, Ending, section headers), whereas DocPres and KCTV have no such mechanism. The coherence score therefore appears to measure, in part, exactly what PPTAgent contributes, making the largest reported improvement an artifact of the evaluation design. I request a human side-by-side comparison between PPTAgent and the baselines, or an alternative coherence evaluation that does not reward the presence of structural slides.
  2. [§5.5, Table 5] The validity evidence for PPTEval is insufficient for the strength of the central claim. The human-LLM Pearson correlation for coherence is 0.55 (Spearman 0.57), the lowest of the three dimensions, and Appendix B reports Fleiss' kappa of 0.54 for coherence, also the lowest. Since the largest PPTAgent advantage over the baselines is in coherence, the evaluator is least reliable exactly on the dimension that carries the main result. The paper does not report a human evaluation comparing PPTAgent directly with DocPres and KCTV, so the judge's bias on the method-comparison question is not calibrated.
  3. [§4.5, Table 3] The claim of "statistically significant" improvements is unsupported by any significance test, confidence interval, or error bar over the 50 presentations per configuration. Moreover, the PPL column shows PPTAgent at 721.54 versus 10–13 for both baselines, a dramatic degradation in the one language-fluency metric reported; the discussion in §4.5 dismisses PPL as inconsistent without a quantitative account. The authors should report paired comparisons (e.g., bootstrap or Wilcoxon over presentations) and explain the PPL discrepancy if the quality claim is to stand.
minor comments (5)
  1. [§4.5, Table 3] In the DocPres/Qwen2.5LM row, the Success Rate is reported as 100.4%, which is impossible for a percentage; this appears to be a typo for 100.0%.
  2. [§4.4] The FID metric is computed using a 64-dimensional output vector, but the feature extractor and the reason for choosing 64 dimensions are not specified; this prevents reproduction of the metric.
  3. [§5.5] The statement that PPTEval's average Pearson correlation of 0.71 "exceeds the scores of other evaluation methods (Kwan et al., 2024)" is misleading, since Kwan et al. is MT-Eval, a multi-turn LLM benchmark, not a presentation-quality evaluation method.
  4. [Appendix C, Table 8] The configuration column in Table 8 reads "Qwen2-VLLM+Qwen2-VLVM", but the text and the rest of the paper describe the best open-source configuration as Qwen2.5LM+Qwen2-VLVM; the label should be corrected.
  5. [Limitations] The Limitations section acknowledges occasional overlapping elements and imperfect use of visual information, which sits in tension with the blanket claim of superior Design scores; a sentence explaining how the design dimension accommodates these failures would help the reader interpret Table 3.

Circularity Check

1 steps flagged · score 6.0 of 10

PPTEval's coherence rubric explicitly rewards the structural slides PPTAgent adds, making the headline coherence gain a designed-in consequence rather than an independent result.

  1. self definitional [Section 3/Table 1; Section 4.5/Table 3; Appendix F (Fig. 20, Fig. 23); Section 5.1/Table 4]
    ""4 Points (Good): The logical flow is reasonable and include basic background information (e.g., speaker or acknowledgments/conclusion)." (Fig. 23); "Identify explicit background information ... such as the author, speaker, date, and other directly stated details, from the opening and closing slides." (Fig. 20); "removing the outline and structural slides significantly degrades coherence (from 4.48 to 3.36/3.45)" (Sec. 5.1)."

    PPTEval's coherence score is not an independent measure of the headline coherence superiority: its top levels are defined by the presence of background/structural slides ("speaker or acknowledgments/conclusion"), and its extraction prompt explicitly locates this information "from the opening and closing slides." PPTAgent's own pipeline is built to create exactly this: Stage I identifies "structural slides (e.g., opening slides)", and outline generation selects "an appropriate reference slide (e.g., opening slide for the first slide)". The baselines have no such structural-slide mechanism. The paper's ablation confirms the link: removing outline or structural slides drops coherence from 4.48 to 3.36/3.45, while content and design stay nearly unchanged.

full rationale

The strongest claim, that PPTAgent significantly outperforms both baselines across Content, Design, and Coherence, is measured entirely by PPTEval, an evaluation framework introduced in the same paper. The Content and Design dimensions have some independent grounding: they reflect general readability and visual-quality criteria, and the human-correlation study (average Pearson 0.71; Design 0.90) provides partial external validation. The Coherence dimension is different in kind. Its scoring rubric explicitly awards 4 points for "basic background information (e.g., speaker or acknowledgments/conclusion)" and 5 points for "detailed and comprehensive background information", and the extraction prompt tells the judge to find this information in the opening and closing slides. PPTAgent's whole pipeline is organized around detecting and reproducing exactly those structural slides: Stage I clusters "structural slides (e.g., opening slides)", and Stage II deliberately chooses an opening slide for the first new slide. The ablation study confirms the sensitivity: removing outline or structural slides drops the coherence score from 4.48 to 3.36 or 3.45, while content/design scores are barely affected. The headline coherence advantage over DocPres and KCTV (4.48 vs 3.24-3.57) is therefore substantially a built-in consequence of the evaluation rubric, not an external confirmation that the generated story is more coherent. The human-correlation data do not break this loop: they validate the judge against human ratings using the same rubric, and no human side-by-side comparison between PPTAgent and the baselines is reported. In addition, the paper states "statistically significant" improvements without reporting significance tests or error bars, so the quantitative force of the claim rests on the self-aligned subscore. This is partial circularity: one of the three headline dimensions reduces to the method's own design choices, even though the Content and Design comparisons retain some independent content. A score of 6 reflects that the central empirical claim relies in part on a metric whose key dimension is confounded with the method's contribution, without claiming the entire derivation is empty.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim depends on hand-chosen thresholds and retry limits, plus domain assumptions about LLM reliability and the validity of the author-designed evaluation. No new physical or conceptual entities are introduced; the invented entities ledger is empty.

free parameters (4)
  • Slide clustering similarity threshold theta = 0.65
    Chosen by hand in Appendix D (Algorithm 1); determines how slides are grouped into layout clusters, which affects reference selection.
  • Image deduplication cosine threshold = 0.85
    Set in Appendix A to remove duplicate images; no tuning analysis provided.
  • Slide deduplication cosine threshold = 0.80
    Set in Appendix A following Fu et al. (2022); affects which slides remain in the reference deck.
  • Maximum self-correction retries = 2
    Set in Section 4.2; limits how many times the LLM can retry edit actions and directly affects success rate.
assumptions (5)
  • domain assumption LLM/MLLM can accurately cluster slides, extract content schemas, and generate correct edit actions from HTML-rendered slides.
    Stage I and Stage II rely on this; prompts in Appendix D and F are the only mechanism enforcing it.
  • domain assumption Reference presentations are high enough quality to transfer design and structure.
    The method copies layouts from references; authors acknowledge in Limitations that output quality depends on the input reference presentation.
  • domain assumption PPTEval's three dimensions (content, design, coherence) adequately capture presentation quality.
    The criteria are grounded in Duarte's books but were selected by the authors; no external benchmark establishes their completeness.
  • domain assumption Four graduate student raters provide a reliable gold standard.
    Appendix B reports Fleiss kappa 0.59, which is moderate agreement; the correlation numbers in Table 5 inherit this limitation.
  • domain assumption GPT-4o judge scores are unbiased across generator models.
    PPTEval uses GPT-4o as judge on presentations generated by GPT-4o, Qwen2.5, and Qwen2-VL; no bias analysis vs. generator identity is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PPTAgent: Generating and Evaluating Presentations Beyond Text-to-Slides." pith.science (2026). https://pith.science/paper/BFDN4W6J

@misc{pith2026250103936,
  author       = {Pith},
  title        = {Pith review of: PPTAgent: Generating and Evaluating Presentations Beyond Text-to-Slides},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BFDN4W6J}},
  note         = {Machine review of arXiv:2501.03936}
}
read the original abstract

Automatically generating presentations from documents is a challenging task that requires accommodating content quality, visual appeal, and structural coherence. Existing methods primarily focus on improving and evaluating the content quality in isolation, overlooking visual appeal and structural coherence, which limits their practical applicability. To address these limitations, we propose PPTAgent, which comprehensively improves presentation generation through a two-stage, edit-based approach inspired by human workflows. PPTAgent first analyzes reference presentations to extract slide-level functional types and content schemas, then drafts an outline and iteratively generates editing actions based on selected reference slides to create new slides. To comprehensively evaluate the quality of generated presentations, we further introduce PPTEval, an evaluation framework that assesses presentations across three dimensions: Content, Design, and Coherence. Results demonstrate that PPTAgent significantly outperforms existing automatic presentation generation methods across all three dimensions.

Figures

Figures reproduced from arXiv: 2501.03936 by the authors.

Figure 1
Figure 1. Comparison between our PPTAGENT ap￾proach (left) and the conventional abstractive summa￾rization method (right). dal et al., 2024) by leveraging the generalization ca￾pabilities of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs). Existing approaches typically follow a text-to￾slides paradigm, which converts LLM outputs into slides using predefined rules or templates. As shown in [PITH_FULL… view at source ↗
Figure 2
Figure 2. Overview of the PPTAGENT workflow. StageI: Presentation Analysis involves analyzing the input presentation to cluster slides into groups and extract their content schemas. Stage II: Presentation Generation generates new presentations guided by the outline, incorporating self-correction mechanisms to ensure robustness. layout and modal complexity of presentations, it is difficult for LLMs to directly determine which … view at source ↗
Figure 3
Figure 3. PPTEVAL assesses presentations from three dimensions: content, design, and coherence. The complete evaluation process is illustrated in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: Score distributions of presentations generated by [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: The number of iterative self-corrections re [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Correlation heatmap between existing auto [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 10
Figure 10. Figure 10: Example of rendering a slide into HTML format [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 9
Figure 9. Figure 9: Example of slide clusters [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 11
Figure 11. Figure 11: The first 60 lines of the XML representation [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Illustration of the prompt used for clustering [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]
Figure 13
Figure 13. Figure 13: Illustration of the prompt used to infer layout [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]
Figure 14
Figure 14. Figure 14: Illustration of the prompt used to extract the [PITH_FULL_IMAGE:figures/full_fig_p015_14.png]
Figure 16
Figure 16. Figure 16: Illustration of the prompt used for generating [PITH_FULL_IMAGE:figures/full_fig_p015_16.png]
Figure 15
Figure 15. Figure 15: Illustration of the prompt used for generating [PITH_FULL_IMAGE:figures/full_fig_p015_15.png]
Figure 18
Figure 18. Figure 18: Illustration of the prompt used to describe [PITH_FULL_IMAGE:figures/full_fig_p016_18.png]
Figure 19
Figure 19. Figure 19: Illustration of the prompt used to describe [PITH_FULL_IMAGE:figures/full_fig_p016_19.png]
Figure 20
Figure 20. Figure 20: Illustration of the prompt used to extract [PITH_FULL_IMAGE:figures/full_fig_p016_20.png]
Figure 23
Figure 23. Figure 23: Illustration of the prompt used to evaluate [PITH_FULL_IMAGE:figures/full_fig_p016_23.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. AGENTIF: Benchmarking Instruction Following of Large Language Models in Agentic Scenarios

    cs.AI 2025-05 conditional novelty 7.0 of 10

    AgentIF introduces a realistic, long-form instruction-following benchmark for agentic scenarios and shows that current LLMs follow fewer than 30% of such instructions perfectly.

  2. OmniPresent: Generating Coherent Presentation Suites from Scientific Papers

    cs.SE 2026-07 conditional novelty 6.5 of 10

    A multi-agent HTML pipeline with shared knowledge and cross-artifact verify-and-repair generates coherent poster/slides/video/page suites from papers and beats specialized baselines on OmniPreBench.

  3. ResearchStudio-Reel: Automate the Last Mile of Research from Paper to Poster, Video, and Blog

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A five-skill agent pipeline with one shared paper extractor and hard render gates produces editable posters, videos, and bilingual blogs, leading the Paper2Poster benchmark on aesthetics.

  4. PosterForest: Hierarchical Multi-Agent Collaboration for Scientific Poster Generation

    cs.AI 2025-08 unverdicted novelty 6.0 of 10

    PosterForest uses a Poster Tree intermediate representation and hierarchical multi-agent reasoning to generate coherent scientific posters without training, outperforming prior methods in evaluations.

  5. AmbiGraph-Eval: Can LLMs Effectively Handle Ambiguous Graph Queries?

    cs.DB 2025-08 unverdicted novelty 6.0 of 10

    A benchmark organized by a six-type taxonomy of ambiguous graph queries reportedly shows that nine LLMs, including top models, frequently produce wrong query translations.

  6. PresentAgent: Multimodal Agent for Presentation Video Generation

    cs.CV 2025-07 reject novelty 5.0 of 10

    PresentAgent chains LLM segmentation, slide rendering, TTS, and ffmpeg to turn documents into narrated presentation videos, but the human-level claim rests on five documents and an unvalidated VLM judge.

  7. SlideCoder: Layout-aware RAG-enhanced Hierarchical Slide Generation from Design

    cs.CV 2025-06 conditional novelty 5.0 of 10

    SlideCoder converts slide design images to editable python-pptx code and reports large gains over prior baselines on a new difficulty-tiered benchmark.

Reference graph

Works this paper leans on

38 extracted references · 12 canonical work pages · cited by 7 Pith papers

  1. [1]

    Sambaran Bandyopadhyay, Himanshu Maheshwari, Anandhavelu Natarajan, and Apoorv Saxena. 2024. Enhancing presentation slide generation by llms with a multi-staged end-to-end approach. arXiv preprint arXiv:2406.06556

    arXiv 2024

  2. [2]

    Andrea Barrick, Dana Davis, and Dana Winkler. 2018. Image versus text in powerpoint lectures: Who does it benefit? Journal of Baccalaureate Social Work, 23(1):91--109

    work page 2018

  3. [3]

    Isabel Alyssa Cachola, Silviu Cucerzan, Allen Herring, Vuksan Mijovic, Erik Oveson, and Sujay Kumar Jauhar. 2024. https://aclanthology.org/2024.findings-emnlp.906 Knowledge-centric templatic views of documents . In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 15460--15476, Miami, Florida, USA. Association for Computational ...

    work page 2024

  4. [4]

    Dongping Chen, Ruoxi Chen, Shilin Zhang, Yinuo Liu, Yaochen Wang, Huichi Zhou, Qihui Zhang, Pan Zhou, Yao Wan, and Lichao Sun. 2024 a . Mllm-as-a-judge: Assessing multimodal llm-as-a-judge with vision-language benchmark. arXiv preprint arXiv:2402.04788

    arXiv 2024

  5. [5]

    Jianlv Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, and Zheng Liu. 2024 b . Bge m3-embedding: Multi-lingual, multi-functionality, multi-granularity text embeddings through self-knowledge distillation. arXiv preprint arXiv:2402.03216

    arXiv 2024

  6. [6]

    Xiang Deng, Yu Gu, Boyuan Zheng, Shijie Chen, Sam Stevens, Boshi Wang, Huan Sun, and Yu Su. 2024. Mind2web: Towards a generalist agent for the web. Advances in Neural Information Processing Systems, 36

    2024

  7. [7]

    Nancy Duarte. 2008. Slide: ology: The art and science of creating great presentations, volume 1. O'Reilly Media Sebastapol

    work page 2008

  8. [8]

    Nancy Duarte. 2010. Resonate: Present visual stories that transform audiences. John Wiley & Sons

    work page 2010

Show all 38 references
  1. [9]

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  2. [10]

    European Organization For Nuclear Research and OpenAIRE . 2013. https://doi.org/10.25495/7GXK-RD71 Zenodo

    2013 doi

  3. [11]

    Weixi Feng, Wanrong Zhu, Tsu-jui Fu, Varun Jampani, Arjun Akula, Xuehai He, Sugato Basu, Xin Eric Wang, and William Yang Wang. 2024. Layoutgpt: Compositional visual planning and generation with large language models. Advances in Neural Information Processing Systems, 36

    2024

  4. [12]

    Tsu-Jui Fu, William Yang Wang, Daniel McDuff, and Yale Song. 2022. https://doi.org/10.1609/aaai.v36i1.19943 Doc2ppt: Automatic presentation slides generation from scientific documents . Proceedings of the AAAI Conference on Artificial Intelligence, 36(1):634--642

    2022 doi

  5. [13]

    Jiaxin Ge, Zora Zhiruo Wang, Xuhui Zhou, Yi-Hao Peng, Sanjay Subramanian, Qinyue Tan, Maarten Sap, Alane Suhr, Daniel Fried, Graham Neubig, et al. 2025. Autopresent: Designing structured visuals from scratch. arXiv preprint arXiv:2501.00912

    2025 arXiv

  6. [14]

    Michael Robert Gryk. 2022. Human readability of data files. Balisage series on markup technologies, 27

    2022

  7. [15]

    Xinyan Guan, Yanjiang Liu, Hongyu Lin, Yaojie Lu, Ben He, Xianpei Han, and Le Sun. 2024. Mitigating large language model hallucinations via autonomous knowledge graph-based retrofitting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 18126--18134

    2024

  8. [16]

    Yiduo Guo, Zekai Zhang, Yaobo Liang, Dongyan Zhao, and Duan Nan. 2023. Pptc benchmark: Evaluating large language models for powerpoint task completion. arXiv preprint arXiv:2311.01767

    2023 arXiv

  9. [17]

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. 2017. Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30

    2017

  10. [18]

    Ryo Kamoi, Yusen Zhang, Nan Zhang, Jiawei Han, and Rui Zhang. 2024. When can llms actually correct their own mistakes? a critical survey of self-correction of llms. Transactions of the Association for Computational Linguistics, 12:1417--1440

    2024

  11. [19]

    Wai-Chung Kwan, Xingshan Zeng, Yuxin Jiang, Yufei Wang, Liangyou Li, Lifeng Shang, Xin Jiang, Qun Liu, and Kam-Fai Wong. 2024. https://arxiv.org/abs/2401.16745 Mt-eval: A multi-turn capabilities evaluation benchmark for large language models . Preprint, arXiv:2401.16745

    2024 arXiv

  12. [20]

    Woosuk Kwon, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph Gonzalez, Hao Zhang, and Ion Stoica. 2023. Efficient memory management for large language model serving with pagedattention. In Proceedings of the 29th Symposium on Operating Systems Princip...

    2023

  13. [21]

    Md Tahmid Rahman Laskar, Sawsan Alqahtani, M Saiful Bari, Mizanur Rahman, Mohammad Abdullah Matin Khan, Haidar Khan, Israt Jahan, Amran Bhuiyan, Chee Wei Tan, Md Rizwan Parvez, Enamul Hoque, Shafiq Joty, and Jimmy Huang. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.764 A ...

    2024 doi

  14. [22]

    Yanda Li, Chi Zhang, Wanqi Yang, Bin Fu, Pei Cheng, Xin Chen, Ling Chen, and Yunchao Wei. 2024. Appagent v2: Advanced agent for flexible mobile interactions. arXiv preprint arXiv:2408.11824

    2024

  15. [23]

    Chin-Yew Lin. 2004. https://aclanthology.org/W04-1013/ ROUGE : A package for automatic evaluation of summaries . In Text Summarization Branches Out, pages 74--81, Barcelona, Spain. Association for Computational Linguistics

    2004

  16. [24]

    Yang Liu, Dan Iter, Yichong Xu, Shuohang Wang, Ruochen Xu, and Chenguang Zhu. 2023. https://doi.org/10.18653/v1/2023.emnlp-main.153 G -eval: NLG evaluation using gpt-4 with better human alignment . In Proceedings of the 2023 Conference on Empirical Methods in Natural Language ...

    2023 doi

  17. [25]

    Himanshu Maheshwari, Sambaran Bandyopadhyay, Aparna Garimella, and Anandhavelu Natarajan. 2024. Presentations are not always linear! gnn meets llm for document-to-presentation transformation with attribution. arXiv preprint arXiv:2405.13095

    2024 arXiv

  18. [26]

    Ishani Mondal, S Shwetha, Anandhavelu Natarajan, Aparna Garimella, Sambaran Bandyopadhyay, and Jordan Boyd-Graber. 2024. Presentations by the humans and for the humans: Harnessing llms for generating persona-aware slides from documents. In Proceedings of the 18th Conference of...

    2024

  19. [27]

    Athar Sefid, Prasenjit Mitra, and Lee Giles. 2021. Slidegen: an abstractive section-based slide generator for scholarly documents. In Proceedings of the 21st ACM Symposium on Document Engineering, pages 1--4

    2021

  20. [28]

    Edward Sun, Yufang Hou, Dakuo Wang, Yunfeng Zhang, and Nancy XR Wang. 2021. D2s: Document-to-slide generation via query-based text summarization. arXiv preprint arXiv:2105.03664

    2021 arXiv

  21. [29]

    Hao Tang, Darren Key, and Kevin Ellis. 2025. Worldcoder, a model-based llm agent: Building world models by writing code and interacting with the environment. Advances in Neural Information Processing Systems, 37:70148--70212

    2025

  22. [30]

    VikParuchuri. 2023. https://github.com/VikParuchuri/marker/ marker

    2023

  23. [31]

    Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. 2024 a . Qwen2-vl: Enhancing vision-language model's perception of the world at any resolution. arXiv preprint arXiv:2409.12191

    2024 arXiv

  24. [32]

    Xingyao Wang, Yangyi Chen, Lifan Yuan, Yizhe Zhang, Yunzhu Li, Hao Peng, and Heng Ji. 2024 b . Executable code actions elicit better llm agents. arXiv preprint arXiv:2402.01030

    2024 arXiv

  25. [33]

    Bichen Wu, Chenfeng Xu, Xiaoliang Dai, Alvin Wan, Peizhao Zhang, Zhicheng Yan, Masayoshi Tomizuka, Joseph Gonzalez, Kurt Keutzer, and Peter Vajda. 2020. https://arxiv.org/abs/2006.03677 Visual transformers: Token-based image representation and processing for computer vision . ...

    2020 arXiv

  26. [34]

    Tong Wu, Guandao Yang, Zhibing Li, Kai Zhang, Ziwei Liu, Leonidas Guibas, Dahua Lin, and Gordon Wetzstein. 2024. Gpt-4v (ision) is a human-aligned evaluator for text-to-3d generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2...

    2024

  27. [35]

    An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, et al. 2024. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115

    2024 arXiv

  28. [36]

    Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena. Advances in Neural Information Processing Systems, 36:46595--46623

    2023

  29. [37]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  30. [38]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.