Autonomous Imagination: Closed-Loop Decomposition of Visual-to-Textual Conversion in Visual Reasoning for Multimodal Large Language Models

TL;DR

This paper introduces autonomous imagination, a method enabling multimodal large language models to iteratively modify visual inputs, decomposing complex visual reasoning tasks into manageable sub-steps without retraining.

Contribution

It proposes a novel closed-loop visual modification approach that enhances MLLMs' reasoning capabilities by decomposing visual-to-textual conversion into iterative steps.

Findings

MLLMs can solve previously unsolvable visual tasks with visual modification.

Closed-loop visual modification improves reasoning without retraining.

The approach is effective across various visual reasoning tasks.

Abstract

Under pure textual modality, Large Language Models (LLMs) have demonstrated remarkable success in complex reasoning tasks by decomposing them into simpler sub-problems. However, Multimodal Large Language Models (MLLMs) still struggle with some seemingly straightforward visual tasks, such as counting and solving jigsaw puzzles. We argue that these tasks challenge the ability of visual-to-textual conversion, where MLLMs convert visual information perceived from the input scene, to textual information for further reasoning and generating the answer. If the complexity of the visual input is beyond the perceptual capability of the MLLMs, without decomposing this conversion process, simply scaling inference-time reasoning cannot solve the task because it repeatedly encounters the same perceptual bottleneck. We propose an approach, autonomous imagination, to enable MLLMs to iteratively modify visual inputs (e.g. isolating objects, rearranging puzzle pieces) into intermediate visual states, decomposing visual-to-textual conversion into closed-loop visual modification steps. We show that, without any retraining, MLLMs can now solve tasks initially beyond their perceptual capability, highlighting that closed-loop visual modification can be an effective way of decomposing the visual reasoning task into solvable substeps. Our code and data are released at https://future-item.github.io/autoimagine-site/.