The core challenge in video understanding lies in perceiving dynamic content changes over time. However, multi-modal large language models (MLLMs) struggle with temporal-sensitive video tasks, such as video temporal grounding, which requires generating timestamps to mark the occurrence of specific events. Existing strategies require MLLMs to generate absolute or relative timestamps directly. We have observed that those MLLMs tend to rely more on language patterns than visual cues when generating timestamps, affecting their performance. To address this problem, we propose VideoExpert, a general-purpose MLLM suitable for several temporal-sensitive video tasks. Inspired by the expert concept, VideoExpert integrates two parallel modules: the Temporal Expert and the Spatial Expert. The Temporal Expert is responsible for modeling time sequences and performing temporal grounding. It processes high-frame-rate yet compressed tokens to capture dynamic variations in videos and includes a lightweight prediction head for precise event localization. The Spatial Expert focuses on content detail analysis and instruction following. It handles specially designed spatial tokens and language input, aiming to generate content-related responses. These two experts collaborate seamlessly via a special token, ensuring coordinated temporal grounding and content generation. Notably, the Temporal and Spatial Experts maintain independent parameter sets. This parameter decoupling design enables specialized learning within each part without mutual interference. By offloading temporal grounding from content generation, VideoExpert prevents text pattern biases in timestamp predictions. Moreover, we introduce a Spatial Compress module to obtain spatial tokens. This module filters and compresses patch tokens while preserving key information, delivering compact yet detail-rich input for the Spatial Expert. Extensive experiments conducted on four widely-used benchmarks (i.e. Charades-STA, QVHighlight, YouCookII and NextGQA) across four tasks (temporal grounding, highlight detection, dense video captioning and grounding question answering) demonstrate the effectiveness and versatility of the VideoExpert.

Temporal-sensitive video tasks, such as dense video captioning, require models to describe event content in detail and generate timestamps marking their occurrence. Existing approaches tend to over-rely on visuals while neglecting the role of rich contexts, limiting temporal reasoning and semantic comprehension in videos. To this end, we propose Time-IC, a multimodal large language model that incorporates video context for improved understanding. Specifically, Time-IC integrates both internal context (timestamps, speech transcripts) and external context (titles, author descriptions), providing complementary semantic cues and high-level semantic priors. More importantly, it unifies context, video frames, and task instructions into a temporally aligned, interleaved sequence. With instruction tuning, this design enables the model to associate contextual elements and generate comprehensive responses across diverse tasks. Extensive experiments on widely-used benchmarks (YouCookII, QVHighlights) across three tasks (dense video captioning, temporal grounding, highlight detection) demonstrate the effectiveness and flexibility of the proposed Time-IC.

The core challenge in video compositional representation lies in jointly identifying atomic actions (verbs) and their associated objects (nouns), while generalizing to unseen verb-noun combinations. Existing approaches often adopt a shared backbone with multi-head classifiers, leading to semantic entanglement and recognition imbalance, especially under ViT-based paradigms. To tackle these issues, we propose a novel framework, Structural Pre-Decoupling and Co-Aligning (SODA), which structurally decouples the early-stage learning paths of verbs and nouns, mitigating semantic interference and recognition imbalance. A core component is the Divide-and-Conquer Disentanglement (DCD) module, comprising two parallel paths: an Object-Removed Verb (OR-V) path that explicitly suppresses object appearance by integrating frame-difference features for short-term motion and trajectory cues for long-term dynamics, and an Object-Centric Noun (OC-N) path that constructs adaptive key-frame representations in the guidance of dynamic features from OR-V. On top of these, a Compositional Co-Alignment (CCA) strategy is further introduced that aligns atomic and compositional representations in a shared semantic space through contrastive learning, capturing implicit commonsense associations of verb-noun pairs while preserving the discriminative power of each stream. Extensive experiments on both standard and zero-shot compositional action recognition benchmarks validate the effectiveness and generalization of our approach.

Video moment retrieval and highlight detection have received attention in the current era of video content proliferation, aiming to localize moments and estimate clip relevances based on user-specific queries. Most existing methods approach these challenges from a discriminative learning perspective, focusing on learning the correspondence between query and activity boundary locations through complex cross-modal interactions. However, the continuous nature of video content often results in unclear boundaries between temporal events. This boundary ambiguity may confuse models, resulting in the subpar performance in predicting target boundaries. To alleviate this problem, we propose to solve the two tasks jointly from the perspective of denoising generation. Moreover, the target boundary can be localized clearly by iterative refinement from coarse to fine. Specifically, a novel framework, DiffusionVMR, is proposed to redefine the two tasks as a unified conditional denoising generation process by combining the diffusion model. During training, the Gaussian noise is added to corrupt the ground truth (GT), with noisy candidates produced as input. The model is trained to reverse this noise addition process. In the inference phase, DiffusionVMR initiates directly from Gaussian noise and progressively refines the proposals from the noise to the meaningful output. Notably, the proposed DiffusionVMR inherits the advantages of diffusion models that allow for iteratively refined results during inference, enhancing the boundary transition from coarse to fine. Furthermore, the training and inference of DiffusionVMR are decoupled. An arbitrary setting can be used in DiffusionVMR during inference without consistency with the training phase. Extensive experiments conducted on five widely used benchmarks (i.e., QVHighlight, Charades-STA, TACoS, YouTubeHighlights, and TVSum) across two tasks (moment retrieval and/or highlight detection) demonstrate the effectiveness and flexibility of the proposed DiffusionVMR.

To solve the essential objective of LDA-L1, NLDA-L1 proposes a nongreedy algorithm by constructing an auxiliary function. In this correspondence, we show that essentially, this algorithm directly solves the objective using a gradient ascending procedure, meaning that the auxiliary function may be not necessary. Then, we further show that NLDA-L1 is a special case of ILDA-L1, which applies the same iterative procedure of ILDA-L1.

Twin support vector clustering (TWSVC) is a recently proposed powerful k-plane clustering method. It, however, is prone to outliers due to the utilization of squared L2-norm distance. Besides, TWSVC is computationally expensive, attributing to the need of solving a series of constrained quadratic programming problems (CQPPs) in learning each clustering plane. To address these problems, this brief first develops a new k-plane clustering method called L1-norm distance minimization-based robust TWSVC by using robust L1-norm distance. To achieve this objective, we propose a novel iterative algorithm. In each iteration of the algorithm, one CQPP is solved. To speed up the computation of TWSVC and simultaneously inherit the merit of robustness, we further propose Fast RTWSVC and design an effective iterative algorithm to optimize it. Only a system of linear equations needs to be computed in each iteration. These characteristics make our methods more powerful and efficient than TWSVC. We also conduct some insightful analysis on the existence of local minimum and the convergence of the proposed algorithms. Theoretical insights and effectiveness of our methods are further supported by promising experimental results.