ShowTable: Unlocking Creative Table Visualization with Collaborative Reflection and Refinement

2025-12-15

AI Summary

Existing models struggle to simultaneously ensure data fidelity and aesthetic quality in creative table visualization, particularly lacking deep reasoning, structured planning, and precise data-to-visual mapping capabilities. To address this, we propose a progressive self-correcting pipeline that synergistically integrates multimodal large language models (MLLMs) and diffusion models: the MLLM performs multi-step reasoning, error diagnosis, and reflective planning, while the diffusion model executes high-fidelity visual generation—forming a closed “reason-diagnose-correct-generate” loop. We formally define the novel task of *creative table visualization*, introduce the *reflective refinement* paradigm, and present TableVisBench—the first dedicated benchmark comprising 800 diverse instances with five-dimensional evaluation criteria—as well as a three-stage automated data construction pipeline. Our method achieves significant improvements over all baselines on TableVisBench, demonstrating superior data comprehension, cross-modal reasoning, and iterative error correction.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

While existing generation and unified models excel at general image generation, they struggle with tasks requiring deep reasoning, planning, and precise data-to-visual mapping abilities beyond general scenarios. To push beyond the existing limitations, we introduce a new and challenging task: creative table visualization, requiring the model to generate an infographic that faithfully and aesthetically visualizes the data from a given table. To address this challenge, we propose ShowTable, a pipeline that synergizes MLLMs with diffusion models via a progressive self-correcting process. The MLLM acts as the central orchestrator for reasoning the visual plan and judging visual errors to provide refined instructions, the diffusion execute the commands from MLLM, achieving high-fidelity results. To support this task and our pipeline, we introduce three automated data construction pipelines for training different modules. Furthermore, we introduce TableVisBench, a new benchmark with 800 challenging instances across 5 evaluation dimensions, to assess performance on this task. Experiments demonstrate that our pipeline, instantiated with different models, significantly outperforms baselines, highlighting its effective multi-modal reasoning, generation, and error correction capabilities.

Problem

• Generates creative table visualizations from data
• Addresses deep reasoning and precise data-to-visual mapping
• Synergizes MLLMs with diffusion models for high-fidelity results

Innovation

• MLLM orchestrates reasoning and error correction
• Diffusion models execute high-fidelity visual generation
• Progressive self-correcting process synergizes both models