communication

Existing document-to-presentation generation methods largely neglect visual design principles and structural coherence, limiting their practical utility. To address this, we propose an end-to-end, two-stage editing-based generation framework: Stage I leverages large language models to learn document structural patterns and infer design constraints; Stage II employs code-driven atomic editing actions to jointly optimize content selection, layout arrangement, and stylistic consistency across slides. Furthermore, we introduce PPTEval—the first three-dimensional evaluation benchmark covering content accuracy, visual design合理性, and cross-slide coherence—equipped with interpretable, multi-faceted metrics. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art baselines across all three dimensions. The source code, dataset, and evaluation toolkit are fully open-sourced.

Interactive visualization editors lack real-time guidance grounded in visual communication principles, limiting design quality. To address this, we propose the first end-to-end feedback framework integrating large language models (LLMs), domain-specific visualization design guidelines, and image-aware perceptual filters to generate actionable, personalized natural language design recommendations. Our method injects structured domain knowledge via prompt engineering and leverages perceptual filters to extract salient visual metrics—enhancing LLM reasoning reliability and grounding suggestions in perceptual evidence. In a longitudinal multi-day study with 13 designers spanning novice to expert levels, our system significantly improved iterative refinement quality and depth of design reflection, receiving high usability ratings. This work provides the first empirical validation of LLMs’ effectiveness in delivering experience-agnostic, principle-based feedback for visualization design—establishing a novel paradigm for intelligent, adaptive design assistance tools.

Hybrid/remote meetings commonly suffer from prolonged duration and declining engagement, while conventional fixed-length summaries fail to satisfy heterogeneous user needs—such as rapid skimming versus deep retrospective review. To address this, we propose Recap, an LLM-driven dual-track meeting summarization system. Grounded in cognitive science and discourse theory, Recap introduces the first complementary summarization paradigm comprising “key highlights (for overview)” and “structured, hierarchical minutes (for retrospective navigation).” It integrates organizational context (e.g., slide links) with personalized adaptation mechanisms, advancing AI-generated summaries from generic outputs toward seamless workflow integration. Through a high-fidelity prototype and qualitative studies in authentic Microsoft meeting contexts (N=7), we empirically validate the synergistic value of both summary types in collaborative discussion and consensus building. Furthermore, analysis of user editing behaviors (additions, deletions, modifications) reveals critical human-AI alignment gaps, providing empirical grounding for explainable and editable AI meeting summaries.

Visualization practitioners often lack formal training, resulting in significant gaps in design knowledge. Method: This paper presents the first systematic evaluation of large language models (LLMs) for visualization design consultation, employing a dual-path approach: quantitative analysis (multidimensional comparison of ChatGPT and expert responses from VisGuides) and qualitative analysis (double-blind user studies, content coding, and human-AI co-feedback analysis). Results: While ChatGPT rapidly generates diverse design alternatives, it substantially underperforms human experts in deep contextual understanding, visual intent inference, and support for nonlinear interactions. The study proposes a novel “context-enhanced + interaction-guided” paradigm tailored to design feedback, establishing both theoretical foundations and a practical framework for developing trustworthy LLM-assisted visualization design systems.

Existing element-attribute grid representations for graphic design completion tasks struggle to model variable-length, type-heterogeneous, and multimodal (text-image) structures. Method: We propose a unified interleaved multimodal tokenized document model that jointly encodes syntactic and semantic structures of markup languages (e.g., SVG/HTML) alongside variable-size, alpha-channel-aware local image generation. We introduce a specialized image quantizer for efficient transparent-image tokenization and integrate an enhanced code-large language model with an interleaved multimodal sequence architecture. Contribution/Results: Our model achieves significant improvements over baselines on three design completion tasks—missing template attributes, image synthesis, and text generation—demonstrating its effectiveness in jointly modeling structural logic and visual semantics in design documents.

This work addresses the challenge of inconsistent README quality, which stems from varying audiences and usage contexts, and the inability of existing tools to simultaneously accommodate style, content, and contextual appropriateness. The paper proposes LintMe, a novel linter that uniquely integrates programmatic rules with large language model (LLM)-based content understanding. LintMe enables users to define context-sensitive checking rules via a lightweight domain-specific language (DSL), combining programmatic validations—such as link verification—with LLM-driven semantic assessments like terminology recognition. This approach enhances documentation quality while preserving authorial autonomy. A user study (N=11) demonstrates that LintMe is both usable and flexible, significantly outperforming baseline approaches that rely solely on direct LLM usage, and its scalability is further validated through illustrative case studies.

This work addresses the challenge that presentation authoring is often constrained by factors such as time, audience, and communicative intent, yet existing tools offer little proactive support for leveraging these constraints. Through a qualitative user study, the authors propose the first constraint-driven, multi-session presentation authoring (CMPA) framework, reframing constraints from passive limitations into active design drivers. Based on this framework, they developed ReSlide, a prototype system that enables creators to harness constraints explicitly during narrative construction. User studies demonstrate that ReSlide significantly enhances users’ ability to utilize constraints in shaping their presentations and facilitates flexible content reuse across varying constraint conditions. The findings offer a novel interaction paradigm and design implications for next-generation presentation authoring tools.

To address the challenge of significant user preference heterogeneity and insufficient personalization in academic presentation generation, this paper introduces the first conditional slide generation task grounded in implicit preferences—namely, exemplar slide pairs and visual templates. Methodologically, we propose a human-behavior-inspired agent framework that integrates multimodal prompt modeling with example-driven preference learning, and design a Chain-of-Speech mechanism for joint generation of spoken narration and editable, structured slides. Our approach uniquely enables implicit preference distillation and cross-template generalization. Evaluated on the first user-preference-aware benchmark, our method significantly outperforms existing baselines: generated slides better align with users’ stylistic preferences and communicative intent, while the unified speech-slide output natively supports downstream applications such as video-based presentations.

The impact of source citation visualization on user cognition and behavior in conversational search remains unclear. This study conducts a large-scale, crowdsourced A/B experiment (N=394) comparing four presentation modalities—collapsible lists, hover cards, footer lists, and aligned sidebar layouts—using behavioral logs, pre-/post-test knowledge/attitude surveys, and fine-grained interaction analysis. Results show that high-visibility designs (e.g., sidebars) significantly increase source hover rates and shifts in viewpoint agreement but impede initial knowledge acquisition and engagement interest. Although hover cards elevate hovering frequency, they fail to translate into meaningful clicks or learning gains. Critically, this work challenges the “transparency-is-beneficial” assumption, demonstrating that **spatial placement** and **interaction pacing**, rather than mere visibility, are decisive factors governing source utility. These findings provide empirical grounding and a novel design paradigm for building trustworthy, cognitively supportive conversational search interfaces.

This work addresses the lack of standardized documentation and evaluation methodologies in prompt engineering, which hinders the reproducibility and interpretability of complex prompts. To remedy this, the authors propose “Prompt Cards,” a novel framework that adapts the model card concept to prompt engineering by introducing a structured template to explicitly document a prompt’s design objectives, contextual strategies, evaluation protocols, and ethical considerations. Demonstrated through a “wordification” task, the approach integrates natural language generation with qualitative assessment to enable systematic recording and analysis of the entire prompting pipeline. Prompt Cards substantially enhance transparency, reproducibility, and methodological rigor, offering the research community a scalable standard for prompt documentation and a new paradigm for benchmarking beyond conventional metrics.