Multimodal Inconsistency Reasoning (MMIR): A New Benchmark for Multimodal Reasoning Models
This addresses a critical gap for AI researchers and developers by providing a benchmark to evaluate and improve MLLMs' ability to handle inconsistencies in layout-rich multimodal content, though it is incremental as it builds on existing multimodal reasoning work.
The authors tackled the problem of multimodal large language models (MLLMs) struggling with inconsistencies in real-world visual-textual content by introducing the Multimodal Inconsistency Reasoning (MMIR) benchmark, showing that models like o1 outperform others while open-source models remain vulnerable, with models excelling in pairwise inconsistency detection but struggling with single-element errors in complex layouts.
Existing Multimodal Large Language Models (MLLMs) are predominantly trained and tested on consistent visual-textual inputs, leaving open the question of whether they can handle inconsistencies in real-world, layout-rich content. To bridge this gap, we propose the Multimodal Inconsistency Reasoning (MMIR) benchmark to assess MLLMs' ability to detect and reason about semantic mismatches in artifacts such as webpages, presentation slides, and posters. MMIR comprises 534 challenging samples, each containing synthetically injected errors across five reasoning-heavy categories: Factual Contradiction, Identity Misattribution, Contextual Mismatch, Quantitative Discrepancy, and Temporal/Spatial Incoherence. We evaluate six state-of-the-art MLLMs, showing that models with dedicated multimodal reasoning capabilities, such as o1, substantially outperform their counterparts while open-source models remain particularly vulnerable to inconsistency errors. Detailed error analyses further show that models excel in detecting pairwise inconsistencies but struggle with inconsistencies confined to single elements in complex layouts. Probing experiments reveal that single-modality prompting, including Chain-of-Thought (CoT) and Set-of-Mark (SoM) methods, yields marginal gains, revealing a key bottleneck in cross-modal reasoning. Our findings highlight the need for advanced multimodal reasoning and point to future research on multimodal inconsistency.