Visual Abstract Thinking: Enhancing Multimodal Reasoning via Visual Abstraction

ICLR 2026 Conference Submission249 Authors

Keywords: Multimodal Reasoning; Multimodal Large Language Model

TL;DR: We propose a visual abstract thinking paradigm for multimodal reasoning tasks, which is simple but effective and yields good performance across multiple benchmark/

Abstract:

Images usually convey richer detail than text, but often include redundant information, which potentially downgrades multimodal reasoning performance. When faced with lengthy or complex messages, humans tend to employ abstract thinking to convert them into simple and concise abstracts. Inspired by this cognitive strategy, we introduce \textbf{V}isual \textbf{A}bstract \textbf{T}hinking (\textbf{VAT}), a novel paradigm that prompts Multimodal Large Language Models (MLLMs) with visual abstract instead of explicit verbal thoughts or elaborate guidance, permitting a more efficient visual reasoning mechanism via concentrated perception.
VAT encourages models to focus on more essential visual elements, concepts and structural features by undermining redundant information compared with explicit thinking methods, such as Chain-of-thought (CoT) and tool-using approaches, that increase the complexity of reasoning process via inserting verbose intermediate steps and external knowledge.
Experimental results show that VAT consistently empowers different MLLMs in visual perception and reasoning tasks. VAT achieves an average gain of $2.21\%$ over GPT-5 baseline, surpassing the gain of CoT, demonstrating that VAT better enhances multimodal task performance of MLLMs. Additionally, VAT spends fewer tokens while achieving higher performance.
These findings highlight the effectiveness of visual abstract thinking and encourage further exploration of more diverse reasoning paradigms from the perspective of human cognition.

Primary Area: applications to computer vision, audio, language, and other modalities

Submission Number: 249