Skill assessment from video entails rating the quality of a person's physical performance and explaining what could be done better. Today's models specialize for an individual sport, and suffer from the high cost and scarcity of expert-level supervision across the long tail of sports. Towards closing that gap, we explore transferable video representations for skill assessment. Our CROSSTRAINER approach discovers skill-attributes--such as balance, control, and hand positioning--whose meaning transcends the boundaries of any given sport, then trains a multimodal language model to generate actionable feedback for a novel video, e.g., "lift hands more to generate more power" as well as its proficiency level, e.g., early expert. We validate the new model on multiple datasets for both cross-sport (transfer) and intra-sport (in-domain) settings, where it achieves gains up to 60% relative to the state of the art. By abstracting out the shared behaviors indicative of human skill, the proposed video representation generalizes substantially better than an array of existing techniques, enriching today's multimodal large language models.

Multi-domain task incremental learning (MTIL) demands models to master domainspecific expertise while preserving generalization capabilities. Inspired by human lifelong learning [1, 2], which relies on revisiting, aligning, and integrating past experiences, we propose a Learning and Ensembling Bridge Adapters (LEBA) framework. To facilitate cohesive knowledge transfer across domains, specifically, we propose a continuous-domain bridge adaptation module, leveraging the distribution transfer capabilities of Schrödinger bridge for stable progressive learning. To strengthen memory consolidation, we further propose a progressive knowledge ensemble strategy that revisits past task representations via a diffusion model and dynamically integrates historical adapters. For efficiency, LEBA maintains a compact adapter pool through similarity-based selection and employs learnable weights to align replayed samples with current task semantics. Together, these components effectively mitigate catastrophic forgetting and enhance generalization across tasks.

We present LongVPO, a novel two-stage Direct Preference Optimization framework that enables short-context vision-language models to robustly understand ultra-long videos without any long-video annotations. In Stage 1, we synthesize preference triples by anchoring questions to individual short clips, interleaving them with distractors, and applying visual-similarity and question-specificity filtering to mitigate positional bias and ensure unambiguous supervision.

In this work, we propose using a unified representation, termed Factorized Features, for low-level vision tasks, where we test on Single Image Super-Resolution (SISR) and Image Compression. Motivated by the shared principles between these tasks, they require recovering and preserving fine image details, whether by enhancing resolution for SISR or reconstructing compressed data for Image Compression. Unlike previous methods that mainly focus on network architecture, our proposed approach utilizes a basis-coefficient decomposition as well as an explicit formulation of frequencies to capture structural components and multi-scale visual features in images, which addresses the core challenges of both tasks. We replace the representation of prior models from simple feature maps with Factorized Features to validate the potential for broad generalizability. In addition, we further optimize the compression pipeline by leveraging the mergeable-basis property of our Factorized Features, which consolidates shared structures on multiframe compression. Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4% in PSNR over the baseline in Super-Resolution (SR) and 9.35% BD-rate reduction in Image Compression compared to the previous SOTA.

Existing multi-view clustering methods employ various strategies to address datalevel sparsity and view-level dynamic fusion. However, we identify a critical yet overlooked issue: varying sparsity across views. Cross-view sparsity variations lead to encoding discrepancies, heightening sample-level semantic heterogeneity and making view-level dynamic weighting inappropriate. To tackle these challenges, we propose Adaptive Sparse Autoencoders for Multi-View Clustering (SparseMVC), a framework with three key modules. Initially, the sparse autoencoder probes the sparsity of each view and adaptively adjusts encoding formats via an entropymatching loss term, mitigating cross-view inconsistencies. Subsequently, the correlation-informed sample reweighting module employs attention mechanisms to assign weights by capturing correlations between early-fused global and viewspecific features, reducing encoding discrepancies and balancing contributions. Furthermore, the cross-view distribution alignment module aligns feature distributions during the late fusion stage, accommodating datasets with an arbitrary number of views. Extensive experiments demonstrate that SparseMVC achieves state-of-theart clustering performance.

Video anomaly detection (VAD) focuses on identifying anomalies in videos. Supervised methods demand substantial in-domain training data and fail to deliver clear explanations for anomalies. In contrast, training-free methods leverage the knowledge reserves and language interactivity of large pre-trained models to detect anomalies. However, the current fixed-length temporal window sampling approaches struggle to accurately capture anomalies with varying temporal spans. Therefore, we propose VADTree that utilizes a Hierarchical Granularityaware Tree (HGTree) structure for flexible sampling in VAD.

Recent work on latent diffusion models (LDMs) has focused almost exclusively on generative tasks, leaving their potential for discriminative transfer largely unexplored. We introduce Discriminative Vicinity Diffusion (DVD), a novel LDM-based framework for a more practical variant of source-free domain adaptation (SFDA): the source provider may share not only a pre-trained classifier but also an auxiliary latent diffusion module, trained once on the source data and never exposing raw source samples. DVD encodes each source feature's label information into its latent vicinity by fitting a Gaussian prior over its k-nearest neighbors and training the diffusion network to "drift" noisy samples back to label-consistent representations. During adaptation, we sample from each target feature's latent vicinity, apply the frozen diffusion module to generate source-like cues, and use a simple InfoNCE loss to align the target encoder to these cues, explicitly transferring decision boundaries without source access. Across standard SFDA benchmarks, DVD outperforms state-of-the-art methods. We further show that the same latent diffusion module enhances the source classifier's accuracy on in-domain data and boosts performance in supervised classification and domain generalization experiments. DVD thus reinterprets LDMs as practical, privacy-preserving bridges for explicit knowledge transfer, addressing a core challenge in source-free domain adaptation that prior methods have yet to solve.

Robust feature representations are essential for learning-based Multi-View Stereo (MVS), which relies on accurate feature matching. Recent MVS methods leverage Transformers to capture long-range dependencies based on local features extracted by conventional feature pyramid networks. However, the quadratic complexity of Transformer-based MVS methods poses challenges to balance performance and efficiency. Motivated by the global modeling capability and linear complexity of the Mamba architecture, we propose MVSMamba, the first Mamba-based MVS network. MVSMamba enables efficient global feature aggregation with minimal computational overhead. To fully exploit Mamba's potential in MVS, we propose a Dynamic Mamba module (DM-module) based on a novel referencecentered dynamic scanning strategy, which enables: (1) Efficient intra-and interview feature interaction from the reference to source views, (2) Omnidirectional multi-view feature representations, and (3) Multi-scale global feature aggregation. Extensive experimental results demonstrate MVSMamba outperforms state-of-theart MVS methods on the DTU dataset and the Tanks-and-Temples benchmark with both superior performance and efficiency.

Humans understand visual scenes by first capturing a global impression and then refining this understanding into distinct, object-like components. Inspired by this process, we introduce Concept-Guided Self-Supervised Learning (CG-SSL), a novel framework that brings structure and interpretability to representation learning through a curriculum of three training phases: (1) global scene encoding, (2) discovery of visual concepts via tokenised cross-attention, and (3) alignment of these concepts across views. Unlike traditional SSL methods, which simply enforce similarity between multiple augmented views of the same image, CG-SSL accounts for the fact that these views may highlight different parts of an object or scene. To address this, our method establishes explicit correspondences between views and aligns the representations of meaningful image regions. At its core, CG-SSL augments standard SSL with a lightweight decoder that learns and refines concept tokens via cross-attention with patch features. The concept tokens are trained using masked concept distillation and a feature-space reconstruction objective. A final alignment stage enforces view consistency by geometrically matching concept regions under heavy augmentation, enabling more compact, robust, and disentangled representations of scene regions. Across multiple backbone sizes, CGSSL achieves state-of-the-art results on image segmentation benchmarks using kNN and linear probes, substantially outperforming prior methods and approaching, or even surpassing, the performance of leading SSL models trained on over 100 more data. Code and pretrained models will be released.

Event cameras provide asynchronous, low-latency, and high-dynamic-range visual signals, making them ideal for real-time perception tasks such as object detection. However, effectively modeling the temporal dynamics of event streams remains a core challenge. Most existing methods follow frame-based detection paradigms, applying temporal modules only at high-level features, which limits early-stage temporal modeling. Transformer-based approaches introduce global attention to capture long-range dependencies, but often add unnecessary complexity and overlook fine-grained temporal cues. In this paper, we propose a CNN-RNN hybrid framework that rethinks temporal modeling for event-based object detection. Our approach is based on two key insights: (1) introducing recurrent modules at lower spatial scales to preserve detailed temporal information where events are most dense, and (2) utilizing Decoupled Deformable-enhanced Recurrent Layers specifically designed according to the inherent motion characteristics of event cameras to extract multiple spatiotemporal features, and performing independent downsampling at multiple spatiotemporal scales to enable flexible, scale-aware representation learning. These multi-scale features are then fused via a feature pyramid network to produce robust detection outputs. Experiments on Gen1, 1 Mpx and eTram dataset demonstrate that our approach achieves superior accuracy over recent transformer-based models, highlighting the importance of precise temporal feature extraction in early stages. This work offers a new perspective on designing architectures for event-driven vision beyond attention-centric paradigms.