Video Large Language Models (Video LLMs) have made significant advancements in various video understanding tasks. However, long-video scenarios remain challenging due to the tension between limited visual token budgets and the need to capture multiple key events. Existing approaches typically process long videos in two stages, i.e., i) select keyframes and ii) perform detailed perception, which exhibit limitations: they lack a modular mechanism for adaptive capacity allocation and self-correction, resulting in unreliable modeling. To tackle these challenges, we propose MoD-VLLM, a novel Modularized Dynamic-Granularity Video LLM framework for multi-event long video understanding, which unifies temporal grounding and semantic understanding iteratively and self-reflectively. Specifically, we propose a Positive-Negative Video Segments Grounding module and a Modularized Dynamic-Granularity Reflection module, which form a closed loop to progressively localize the question-related video segments. The grounding module instructs a Video LLM to distinguish relevant from irrelevant video segments based on the video question. The reflection module employs a modularized scheduler that dynamically selects fine-grained encoding for relevant positive segments to capture detailed perception and coarse-grained encoding for negative segments to maintain global context. We further propose a dynamic-granularity reinforcement learning strategy, allowing MoD-VLLM to learn optimal grounding policies and dynamic granularity visual representation jointly. Moreover, we propose MEventBench, a challenging Multi-Event Long Video Benchmark for complex long video reasoning. Extensive experiments on several long video understanding benchmarks and our MEventBench demonstrate that MoD-VLLM significantly outperforms state-of-the-art baselines.
Wei Feng, Xin Wang, Yuwei Zhan et al.· 0 citations
Dynamic text-attributed graphs (DyTAGs) exhibit coupled textual and structural dynamics, and existing mainstream approaches for DyTAGs extend conventional large language models (LLMs) to capture both dynamics, thereby giving rise to dynamic graph LLMs. However, in DyTAGs, the continuous emergence of new nodes and edges with incoming textual content and interactions drives the joint evolution of graph structural-textual patterns, causing existing methods to struggle with evolving patterns. This motivates a largely unexplored problem of continual learning on DyTAGs, which aims to adapt to constantly evolving graph structural-textual patterns while retaining past knowledge, which imposes two challenges: 1) unlike common graphs, graph structure and textual semantics in emerging DyTAG patterns jointly evolve, requiring dynamic graph LLMs to adapt structure, text, and graph-text fusion simultaneously; and 2) updating dynamic graph LLMs to fit a new pattern may destroy the global graph-text fusion capabilities and bias the model towards recent local dynamics. To address these challenges, we propose a novel Continual Learning Dynamic Graph LLM framework (Continual-GraphLLM) to continually adapt to incoming patterns by routing them to experts specialized in similar past patterns, while mitigating the overwriting of previously learned patterns by assigning new experts to unseen patterns. Specifically, we propose a graph-text factor-based router to adapt to incoming structural-textual joint patterns by utilizing latent factors to adaptively activate suitable experts. Furthermore, we design invariance regularized multi-scale experts that mitigate forgetting by capturing the invariances among learned patterns assigned to the same expert, where each expert progressively integrates structural and textual information from local scale to global scale. Extensive experiments on real-world DyTAGs demonstrate the superiority of our method over competitive baselines, highlighting its effectiveness in adapting to emerging DyTAG patterns.
Tianhang Wan, Xin Wang, Haibo Chen et al.· Proceedings of the 32nd ACM...· 0 citations
Graph machine learning has witnessed rapid progress across both academia and industry. However, most existing methods are developed under the in-distribution (I.D.) hypothesis, which assumes that training and testing graph data are drawn from the same distribution. In real-world applications—ranging from dynamic knowledge graphs to evolving biomedical networks—this assumption is frequently violated, resulting in severe performance degradation under distribution shifts. Addressing this challenge has become a key focus in recent years, leading to the development of novel paradigms that move beyond the I.D. setting. This tutorial presents a comprehensive overview of three emerging and synergistic directions for tackling distribution shifts in graph learning. First, we highlight Graph LLMs, which combine the representational power of large language models with graph structures to enable flexible, in-context, and few-shot learning on graphs. Second, we introduce adaptation techniques for both GNNs and Graph LLMs, including graph neural architecture search and continual learning strategies for evolving data. Third, we cover generalization methods that incorporate causality and invariance principles to build robust graph models under unseen distributions. We will advocate novel, high-quality research findings, as well as innovative solutions to the challenging problems in graph machine learning under distribution shifts and the applications on graphs. This topic is at the core of the scope of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, and is attractive to machine learning as well as data mining audience from both academia and industry.
Xin Wang, Haoyang Li, Haibo Chen et al.· Proceedings of the 32nd ACM...· 0 citations