Multimodal Instruction Following serves as a fundamental capability of multimodal language models, involving accurate comprehension and execution of user-provided instructions. However, existing multimodal instruction-following datasets and benchmarks face the shortcomings outlined below: (a) Lack of Difficulty Stratification, they collect diverse instruction categories but neglect the stratification of difficulty levels across these categories, which leads to overlap, bias, and low interpretability. (b) Lack of Fine-Grained Metrics, they conflate the model's ability to ``solve tasks" and ``follow constraints" into a single metric, which fails to accurately reflect its instruction-following capability. (c) Lack of Multi-Task Instructions, they overlook the fact that real-world user instructions often consist of multiple combined tasks. This paper proposes MMIFEvol, a framework for multimodal instruction evolving and benchmarking. First, we define the essential components of a carefully curated multimodal instruction set and establish corresponding difficulty levels, based on which we synthesize diverse instruction data. Next, we decouple the evaluation criteria for the instruction following into three different metrics to construct a high-quality benchmark and assess existing models. Experimental results demonstrate that current models still struggle with following complex instructions, while fine-tuning using MMIFEvol data effectively improves models' responsiveness to multimodal instructions.

Multimodal Large Language Models (MLLMs) have recently achieved strong performance across a variety of multimodal tasks. However, they still suffer from various forms of hallucination, which hinder their practical deployment. Prior approaches often struggle to efficiently construct high-quality hallucination-related samples and to process them in a fine-grained manner, resulting in limited effectiveness in hallucination alleviation. To address this issue, we propose a data sampling strategy that selects samples better suited for hallucination-oriented training, thereby enhancing training effectiveness. In addition, we introduce a quantitative method for measuring hallucination severity and assign individualized weights to training samples accordingly. Building on this, we present Hallucination-Differentiated Direct Preference Optimization (HD-DPO), a novel preference optimization framework. During fine-tuning, HD-DPO incorporates these weights into both the formulation of customized loss functions and the modulation of localized visual attention, enabling fine-grained optimization. Experimental results demonstrate that our method outperforms existing fine-tuning strategies across multiple benchmarks and generalizes well to diverse MLLM architectures, effectively reducing hallucination rates and enhancing overall model performance.

Accurate interpretation and visualization of human instructions are crucial for text-to-image (T2I) synthesis. However, current models struggle to capture semantic variations from word order changes, and existing evaluations, relying on indirect metrics like text-image similarity, fail to reliably assess these challenges. This often obscures poor performance on complex or uncommon linguistic patterns by the focus on frequent word combinations. To address these deficiencies, we propose a novel metric called SemVarEffect and a benchmark named SemVarBench, designed to evaluate the causality between semantic variations in inputs and outputs in T2I synthesis. Semantic variations are achieved through two types of linguistic permutations, while avoiding easily predictable literal variations. Experiments reveal that the CogView-3-Plus and Ideogram 2 performed the best, achieving a score of 0.2/1. Semantic variations in object relations are less understood than attributes, scoring 0.07/1 compared to 0.17-0.19/1. We found that cross-modal alignment in UNet or Transformers plays a crucial role in handling semantic variations, a factor previously overlooked by a focus on textual encoders. Our work establishes an effective evaluation framework that advances the T2I synthesis community's exploration of human instruction understanding. Our benchmark and code are available at https://github.com/zhuxiangru/SemVarBench.

Scene understanding enables intelligent agents to interpret and comprehend their environment. While existing large vision-language models (LVLMs) for scene understanding have primarily focused on indoor household tasks, they face two significant limitations when applied to outdoor large-scale scene understanding. First, outdoor scenarios typically encompass larger-scale environments observed through various sensors from multiple viewpoints (e.g., bird view and terrestrial view), while existing indoor LVLMs mainly analyze single visual modalities within building-scale contexts from humanoid viewpoints. Second, existing LVLMs suffer from missing multidomain perception outdoor data and struggle to effectively integrate 2D and 3D visual information. To address the aforementioned limitations, we build the first multidomain perception outdoor scene understanding dataset, named SVM-City, deriving from multi-Scale scenarios with multi-View and multi-Modal instruction tuning data. It contains 420k images and 4, 811M point clouds with 567k question-answering pairs from vehicles, low-altitude drones, high-altitude aerial planes, and satellite. To effectively fuse multimodal data in the absence of one modality, we introduce incomplete multimodal learning to model outdoor scene understanding and design the LVLM named City-VLM. Multimodal fusion is realized by constructed as a joint probabilistic distribution space rather than implementing directly explicit fusion operations (e.g., concatenation). Experimental results on three typical outdoor scene understanding tasks show City-VLM achieves 18.14 % performance surpassing existing LVLMs in question-answering tasks averagely. Our method demonstrates pragmatic and generalization performance across multiple outdoor scenes.

Multimodal knowledge bases (MMKBs) provide cross-modal aligned knowledge crucial for multimodal tasks. However, the images in existing MMKBs are generally collected for entities in encyclopedia knowledge graphs. Therefore, detailed groundings of visual semantics with linguistic concepts are lacking, which are essential for the visual concept cognition ability of multimodal models. Addressing this gap, we introduce M2 ConceptBase, the first concept-centric MMKB. M2 ConceptBase models concepts as nodes with associated images and detailed textual descriptions. We propose a context-aware multimodal symbol grounding approach to align concept-image and concept-description pairs using context information from image-text datasets. Comprising 951K images and 152K concepts, M2 ConceptBase links each concept to an average of 6.27 images and a single description, ensuring comprehensive visual and textual semantics. Human studies confirm more than 95% alignment accuracy, underscoring its quality. Additionally, our experiments demonstrate that M2 ConceptBase significantly enhances VQA model performance on the OK-VQA task. M2 ConceptBase also substantially improves the fine-grained concept understanding capabilities of multimodal large language models through retrieval augmentation in two concept-related tasks, highlighting its value.

Recent years have witnessed the resurgence of knowledge engineering which is featured by the fast growth of knowledge graphs. However, most of existing knowledge graphs are represented with pure symbols, which hurts the machine's capability to understand the real world. The multi-modalization of knowledge graphs is an inevitable key step towards the realization of human-level machine intelligence. The results of this endeavor are Multi-modal Knowledge Graphs (MMKGs). In this survey on MMKGs constructed by texts and images, we first give definitions of MMKGs, followed with the preliminaries on multi-modal tasks and techniques. We then systematically review the challenges, progresses and opportunities on the construction and application of MMKGs respectively, with detailed analyses of the strengths and weaknesses of different solutions. We finalize this survey with open research problems relevant to MMKGs.

Image-text retrieval is a challenging cross-modal task that arouses much attention. While the traditional methods cannot break down the barriers between different modalities, Vision-Language Pre-trained (VLP) models greatly improve image-text retrieval performance based on massive image-text pairs. Nonetheless, the VLP-based methods are still prone to produce retrieval results that cannot be cross-modal aligned with entities. Recent efforts try to fix this problem at the pre-training stage, which is not only expensive but also unpractical due to the unavailable of full datasets. In this paper, we novelly propose a lightweight and practical approach to align cross-modal entities for image-text retrieval upon VLP models only at the fine-tuning and re-ranking stages. We employ external knowledge and tools to construct extra fine-grained image-text pairs, and then emphasize cross-modal entity alignment through contrastive learning and entity-level mask modeling in fine-tuning. Besides, two re-ranking strategies are proposed, including one specially designed for zero-shot scenarios. Extensive experiments with several VLP models on multiple Chinese and English datasets show that our approach achieves state-of-the-art results in nearly all settings.