METAL: A Multi-Agent Framework for Chart Generation with Test-Time Scaling

TL;DR

METAL is a multi-agent vision-language framework that improves automatic chart generation by decomposing tasks, enabling test-time scaling, and enhancing self-correction, leading to significant performance gains.

Contribution

The paper introduces METAL, a novel multi-agent framework that decomposes chart generation into collaborative agents and demonstrates test-time scaling for improved results.

Findings

Achieves 5.2% better performance than previous methods.

Performance improves monotonically with increased computational budget.

Separating modalities enhances self-correction in multimodal VLMs.

Abstract

Chart generation aims to generate code to produce charts satisfying the desired visual properties, e.g., texts, layout, color, and type. It has great potential to empower the automatic professional report generation in financial analysis, research presentation, education, and healthcare. In this work, we build a vision-language model (VLM) based multi-agent framework for effective automatic chart generation. Generating high-quality charts requires both strong visual design skills and precise coding capabilities that embed the desired visual properties into code. Such a complex multi-modal reasoning process is difficult for direct prompting of VLMs. To resolve these challenges, we propose METAL, a multi-agent framework that decomposes the task of chart generation into the iterative collaboration among specialized agents. METAL achieves 5.2% improvement over the current best result in the chart generation task. The METAL framework exhibits the phenomenon of test-time scaling: its performance increases monotonically as the logarithmic computational budget grows from 512 to 8192 tokens. In addition, we find that separating different modalities during the critique process of METAL boosts the self-correction capability of VLMs in the multimodal context.