Instruction-Guided Visual Masking

TL;DR

This paper introduces Instruction-guided Visual Masking (IVM), a versatile visual grounding approach that improves multimodal instruction following by focusing on task-relevant image regions, leading to state-of-the-art results.

Contribution

The paper presents a novel visual masking technique and a large dataset, along with a new training method, enhancing multimodal models' ability to align instructions with image regions.

Findings

IVM significantly improves performance on VQA tasks.

IVM enhances robotic control accuracy.

State-of-the-art results achieved across benchmarks.

Abstract

Instruction following is crucial in contemporary LLM. However, when extended to multimodal setting, it often suffers from misalignment between specific textual instruction and targeted local region of an image. To achieve more accurate and nuanced multimodal instruction following, we introduce Instruction-guided Visual Masking (IVM), a new versatile visual grounding model that is compatible with diverse multimodal models, such as LMM and robot model. By constructing visual masks for instruction-irrelevant regions, IVM-enhanced multimodal models can effectively focus on task-relevant image regions to better align with complex instructions. Specifically, we design a visual masking data generation pipeline and create an IVM-Mix-1M dataset with 1 million image-instruction pairs. We further introduce a new learning technique, Discriminator Weighted Supervised Learning (DWSL) for preferential IVM training that prioritizes high-quality data samples. Experimental results on generic multimodal tasks such as VQA and embodied robotic control demonstrate the versatility of IVM, which as a plug-and-play tool, significantly boosts the performance of diverse multimodal models, yielding new state-of-the-art results across challenging multimodal benchmarks. Code, model and data are available at https://github.com/2toinf/IVM.