视听学习

3D spatial reasoning in dynamic, audio-visual environments is a cornerstone of human cognition yet remains largely unexplored by existing Audio-Visual Large Language Models (AV-LLMs) and benchmarks, which predominantly focus on static or 2D scenes. We introduce SAVVY-Bench, the first benchmark for 3D spatial reasoning in dynamic scenes with synchronized spatial audio. SAVVY-Bench is comprised of thousands of carefully curated question–answer pairs probing both directional and distance relationships involving static and moving objects, and requires fine-grained temporal grounding, consistent 3D localization, and multi-modal annotation. To tackle this challenge, we propose SAVVY, a novel training-free reasoning pipeline that consists of two stages: (i) Egocentric Spatial Tracks Estimation, which leverages AV-LLMs as well as other audio-visual methods to track the trajectories of key objects related to the query using both visual and spatial audio cues, and (ii) Dynamic Global Map Construction, which aggregates multi-modal queried object trajectories and converts them into a unified global dynamic map. Using the constructed map, a final QA answer is obtained through a coordinate transformation that aligns the global map with the queried viewpoint. Empirical evaluation demonstrates that SAVVY substantially enhances performance of state-of-the-art AV-LLMs, setting a new standard and stage for approaching dynamic 3D spatial reasoning in AV-LLMs.

Despite their powerful capabilities, multimodal large language models (MLLMs) suffer from considerable computational overhead due to their reliance on massive visual tokens. Recent studies have explored token pruning to alleviate this problem, which typically uses text-vision cross-attention or [CLS] attention to assess and discard redundant visual tokens. In this work, we identify a critical limitation of such attention-first pruning approaches, i.e., they tend to preserve semantically similar tokens, resulting in pronounced performance drops under high pruning rates. To this end, we propose HoloV, a simple yet effective, plug-and-play visual token pruning framework for efficient inference. Distinct from previous attention-first schemes, HoloV rethinks token retention from a holistic perspective. By adaptively distributing the pruning budget across different spatial crops, HoloV ensures that the retained tokens capture the global visual context rather than isolated salient features. This strategy minimizes representational collapse and maintains task-relevant information even under aggressive pruning. Experimental results demonstrate that our HoloV achieves superior performance across various tasks, MLLM architectures, and pruning ratios compared to SOTA methods. For instance, LLaVA1.5 equipped with HoloV preserves 95.8% of the original performance after pruning 88.9% of visual tokens, achieving superior efficiency-accuracy trade-offs.

Despite widespread adoption, multimodal large language models (MLLMs) suffer performance degradation when encountering unfamiliar queries under distribution shifts. Existing methods to improve MLLM generalization typically require either more instruction data or larger advanced model architectures, both of which incur non-trivial human labor or computational costs. In this work, we take an alternative approach to enhance the generalization and robustness of MLLMs under distribution shifts, from a representation learning perspective. Inspired by information bottleneck (IB) principle, we derive a variational lower bound of the IB for MLLMs and devise a practical implementation, Visual Instruction Bottleneck Tuning (Vittle). We then provide a theoretical justification of Vittle by revealing its connection to an information-theoretic robustness metric of MLLM. Empirical validation of multiple MLLMs on open-ended and closed-form question answering and object hallucination detection tasks over 45 datasets, including 30 shift scenarios, demonstrates that Vittle consistently improves the MLLM's robustness under shifts by pursuing the learning of a minimal sufficient representation.

Human-Object Interaction Detection (HOI-DET) aims to localize human-object pairs and identify their interactive relationships. To aggregate contextual cues, existing methods typically propagate information across all detected entities via self‑attention mechanisms, or establish message passing between humans and objects with bipartite graphs. However, they primarily focus on pairwise relationships, overlooking that interactions in real-world scenarios often emerge from collective behaviors ($\textit{i}.\textit{e}.$, multiple humans and objects engaging in joint activities). In light of this, we revisit relation modeling from a $\textit{group}$ view and propose GroupHOI, a framework that propagates contextual information in terms of $\textit{geometric proximity}$ and $\textit{semantic similarity}$. To exploit the geometric proximity, humans and objects are grouped into distinct clusters using a learnable proximity estimator based on spatial features derived from bounding boxes. In each group, a soft correspondence is computed via self-attention to aggregate and dispatch contextual cues. To incorporate the semantic similarity, we enhance the vanilla transformer-based interaction decoder with local contextual cues from HO-pair features. Extensive experiments on HICO-DET and V-COCO benchmarks demonstrate the superiority of GroupHOI over the state-of-the-art methods. It also exhibits leading performance on the more challenging Nonverbal Interaction Detection (NVI-DET) task, which involves varied forms of higher-order interactions within groups.

Understanding long video content is a complex endeavor that often relies on densely sampled frame captions or end-to-end feature selectors, yet these techniques commonly overlook the logical relationships between textual queries and visual elements. In practice, computational constraints necessitate coarse frame subsampling, a challenge analogous to “finding a needle in a haystack.” To address this issue, we introduce a semantics-driven search framework that reformulates keyframe selection under the paradigm of Visual Semantic-Logical Search (VSLS). Specifically, we systematically define four fundamental logical dependencies: 1) spatial co-occurrence, 2) temporal proximity, 3) attribute dependency, and 4) causal order. These relations dynamically update frame sampling distributions through an iterative refinement process, enabling context-aware identification of semantically critical frames tailored to specific query requirements. Our method establishes new state-of-the-art performance on the manually annotated benchmark in keyframe selection metrics. Furthermore, when applied to downstream video question-answering tasks, the proposed approach demonstrates the best performance gains over existing methods on LongVideoBench and Video-MME, validating its effectiveness in bridging the logical gap between textual queries and visual-temporal reasoning. The code will be publicly available.

This paper proposes a novel scene understanding task called Visual Jenga. Drawing inspiration from the game Jenga, the proposed task involves progressively removing objects from a single image until only the background remains. Just as Jenga players must understand structural dependencies to maintain tower stability, our task reveals the intrinsic relationships between scene elements by systematically exploring which objects can be removed while preserving scene coherence in both physical and geometric sense. As a starting point for tackling the Visual Jenga task, we propose a simple, data-driven, training-free approach that is surprisingly effective on a range of real-world images. The principle behind our approach is to utilize the asymmetry in the pairwise relationships between objects within a scene and employ a large inpainting model to generate a set of counterfactuals to quantify the asymmetry.

Vision Transformers rely on fixed patch tokens that ignore the spatial and semantic structure of images. In this work, we introduce an end-to-end differentiable tokenizer that adapts to image content with pixel-level granularity while remaining backward-compatible with existing architectures for retrofitting pretrained models. Our method uses hierarchical model selection with information criteria to provide competitive performance in both image-level classification and dense-prediction tasks, and even supports out-of-the-box raster-to-vector conversion.

Seeking high-quality representations with latent variable models (LVMs) to reveal the intrinsic correlation between neural activity and behavior or sensory stimuli has attracted much interest. In the study of the biological visual system, naturalistic visual stimuli are inherently high-dimensional and time-dependent, leading to intricate dynamics within visual neural activity. However, most work on LVMs has not explicitly considered neural temporal relationships. To cope with such conditions, we propose Time-Evolving Visual Dynamical System (TE-ViDS), a sequential LVM that decomposes neural activity into low-dimensional latent representations that evolve over time. To better align the model with the characteristics of visual neural activity, we split latent representations into two parts and apply contrastive learning to shape them. Extensive experiments on synthetic datasets and real neural datasets from the mouse visual cortex demonstrate that TE-ViDS achieves the best decoding performance on naturalistic scenes/movies, extracts interpretable latent trajectories that uncover clear underlying neural dynamics, and provides new insights into differences in visual information processing between subjects and between cortical regions. In summary, TE-ViDS is markedly competent in extracting stimulus-relevant embeddings from visual neural activity and contributes to the understanding of visual processing mechanisms. Our codes are available at https://github.com/Grasshlw/Time-Evolving-Visual-Dynamical-System.

Spiking Neural Networks (SNNs) demonstrate significant potential for energy-efficient neuromorphic computing through an event-driven paradigm. While training methods and computational models have greatly advanced, SNNs struggle to achieve competitive performance in visual long-sequence modeling tasks. In artificial neural networks, the effective receptive field (ERF) serves as a valuable tool for analyzing feature extraction capabilities in visual long-sequence modeling. Inspired by this, we introduce the Spatio-Temporal Effective Receptive Field (ST-ERF) to analyze the ERF distributions across various Transformer-based SNNs. Based on the proposed ST-ERF, we reveal that these models suffer from establishing a robust global ST-ERF, thereby limiting their visual feature modeling capabilities. To overcome this issue, we propose two novel channel-mixer architectures: \underline{m}ulti-\underline{l}ayer-\underline{p}erceptron-based m\underline{ixer} (MLPixer) and \underline{s}plash-and-\underline{r}econstruct \underline{b}lock (SRB). These architectures enhance global spatial ERF through all timesteps in early network stages of Transformer-based SNNs, improving performance on challenging visual long-sequence modeling tasks. Extensive experiments conducted on the Meta-SDT variants and across object detection and semantic segmentation tasks further validate the effectiveness of our proposed method. Beyond these specific applications, we believe the proposed ST-ERF framework can provide valuable insights for designing and optimizing SNN architectures across a broader range of tasks. The code is available at \href{https://github.com/EricZhang1412/Spatial-temporal-ERF}{\faGithub~EricZhang1412/Spatial-temporal-ERF}.

We study whether visual embedding models capture continuous, ordinal attributes along linear directions, which we term _rank axes_. We define a model as _rankable_ for an attribute if projecting embeddings onto such an axis preserves the attribute's order. Across 7 popular encoders and 9 datasets with attributes like age, crowd count, head pose, aesthetics, and recency, we find that many embeddings are inherently rankable. Surprisingly, a small number of samples, or even just two extreme examples, often suffice to recover meaningful rank axes, without full-scale supervision. These findings open up new use cases for image ranking in vector databases and motivate further study into the structure and learning of rankable embeddings.

Pre-trained stable diffusion models (SD) have shown great advances in visual correspondence. In this paper, we investigate the capabilities of Diffusion Transformers (DiTs) for accurate dense correspondence. Distinct from SD, DiTs exhibit a critical phenomenon in which very few feature activations exhibit significantly larger values than others, known as massive activations, leading to uninformative representations and significant performance degradation for DiTs. The massive activations consistently concentrate at very few fixed dimensions across all image patch tokens, holding little local information. We analyze these dimension-concentrated massive activations and uncover that their concentration is inherently linked to the Adaptive Layer Normalization (AdaLN) in DiTs. Building on these findings, we propose the Diffusion Transformer Feature (DiTF), a training-free AdaLN-based framework that extracts semantically discriminative features from DiTs. Specifically, DiTF leverages AdaLN to adaptively localize and normalize massive activations through channel-wise modulation. Furthermore, a channel discard strategy is introduced to mitigate the adverse effects of massive activations. Experimental results demonstrate that our DiTF outperforms both DINO and SD-based models and establishes a new state-of-the-art performance for DiTs in different visual correspondence tasks (e.g., with +9.4\% on Spair-71k and +4.4\% on AP-10K-C.S.).

Scaling visual generation models is essential for real-world content creation, yet requires substantial training and computational expenses. Alternatively, test-time scaling has garnered growing attention due to resource efficiency and promising performance. In this work, we present the first general test-time scaling framework for visual auto-regressive (VAR) models, TTS-VAR, modeling the generation process as a path searching problem. Inspired by VAR's hierarchical coarse-to-fine multi-scale generation, our framework integrates two key components: (i) At coarse scales, we observe that generated tokens are hard for evaluation, possibly leading to erroneous acceptance of inferior samples or rejection of superior samples. Noticing that the coarse scales contain sufficient structural information, we propose clustering-based diversity search. It preserves structural variety through semantic feature clustering, enabling later selection on samples with higher potential. (ii) In fine scales, resampling-based potential selection prioritizes promising candidates using potential scores, which are defined as reward functions incorporating multi-scale generation history. To dynamically balance computational efficiency with exploration capacity, we further introduce an adaptive descending batch size schedule throughout the causal generation process. Experiments on the powerful VAR model Infinity2B show a notable 8.7% GenEval score improvement (0.69→0.75). Key insights reveal that early-stage structural features effectively influence final quality, and resampling efficacy varies across generation scales.

Transformer-based algorithms, such as LoRAT, have significantly enhanced object-tracking performance. However, these approaches rely on a standard attention mechanism, which incurs quadratic token complexity, making real-time inference computationally expensive. In this paper, we introduce LoRATv2, a novel tracking framework that addresses these limitations with three main contributions. First, LoRATv2 integrates frame-wise causal attention, which ensures full self-attention within each frame while enabling causal dependencies across frames, significantly reducing computational overhead. Moreover, key-value (KV) caching is employed to efficiently reuse past embeddings for further speedup. Second, building on LoRAT's parameter-efficient fine-tuning, we propose Stream-Specific LoRA Adapters (SSLA). As frame-wise causal attention introduces asymmetry in how streams access temporal information, SSLA assigns dedicated LoRA modules to the template and each search stream, with the main ViT backbone remaining frozen. This allows specialized adaptation for each stream's role in temporal tracking. Third, we introduce a two-phase progressive training strategy, which first trains a single-search-frame tracker and then gradually extends it to multi-search-frame inputs by introducing additional LoRA modules. This curriculum-based learning paradigm improves long-term tracking while maintaining training efficiency. In extensive experiments on multiple benchmarks, LoRATv2 achieves state-of-the-art performance, substantially improved efficiency, and a superior performance-to-FLOPs ratio over state-of-the-art trackers. The code is available at https://github.com/LitingLin/LoRATv2.

Multimodal Large Language Models (MLLMs) typically process a large number of visual tokens, leading to considerable computational overhead, even though many of these tokens are redundant. Existing visual token pruning methods primarily focus on selecting the most salient tokens based on attention scores, resulting in the semantic incompleteness of the selected tokens. In this paper, we propose a novel visual token pruning strategy, called **S**aliency-**C**overage **O**riented token **P**runing for **E**fficient MLLMs (SCOPE), to jointly model both the saliency and coverage of the selected visual tokens to better preserve semantic completeness. Specifically, we introduce a set-coverage for a given set of selected tokens, computed based on the token relationships. We then define a token-coverage gain for each unselected token, quantifying how much additional coverage would be obtained by including it. By integrating the saliency score into the token-coverage gain, we propose our SCOPE score and iteratively select the token with the highest SCOPE score. We conduct extensive experiments on multiple vision-language understanding benchmarks using the LLaVA-1.5 and LLaVA-Next models. Experimental results demonstrate that our method consistently outperforms prior approaches.

Contrastive Language-Image Pre-training (CLIP) has become a foundation model and has been applied to various vision and multimodal tasks. However, recent works indicate that CLIP falls short in distinguishing detailed differences in images and shows suboptimal performance on dense-prediction and vision-centric multimodal tasks. Therefore, this work focuses on improving existing CLIP models, aiming to capture as many visual details in images as possible. We find that a specific type of generative models, unCLIP, provides a suitable framework for achieving our goal. Specifically, unCLIP trains an image generator conditioned on the CLIP image embedding. In other words, it inverts the CLIP image encoder. Compared to discriminative models like CLIP, generative models are better at capturing image details because they are trained to learn the data distribution of images. Additionally, the conditional input space of unCLIP aligns with CLIP's original image-text embedding space. Therefore, we propose to invert unCLIP (dubbed un$^2$CLIP) to improve the CLIP model. In this way, the improved image encoder can gain unCLIP's visual detail capturing ability while preserving its alignment with the original text encoder simultaneously. We evaluate our improved CLIP across various tasks to which CLIP has been applied, including the challenging MMVP-VLM benchmark, the dense-prediction open-vocabulary segmentation task, and multimodal large language model tasks. Experiments show that un$^2$CLIP significantly improves the original CLIP and previous CLIP improvement methods. Code and models are available at https://github.com/LiYinqi/un2CLIP.

Audio-visual event parsing plays a crucial role in understanding multimodal video content, but existing methods typically rely on offline processing of entire videos with huge model sizes, limiting their real-time applicability. We introduce Online Audio-Visual Event Parsing (On-AVEP), a novel paradigm for parsing audio, visual, and audio-visual events by sequentially analyzing incoming video streams. The On-AVEP task necessitates models with two key capabilities: (1) Accurate online inference, to effectively distinguish events with unclear and limited context in online settings, and (2) Real-time efficiency, to balance high performance with computational constraints. To cultivate these, we propose the $\textbf{Pre}$dictive $\textbf{F}$uture $\textbf{M}$odeling (PreFM) framework featured by (a) predictive multimodal future modeling to infer and integrate beneficial future audio-visual cues, thereby enhancing contextual understanding and (b) modality-agnostic robust representation along with focal temporal prioritization to improve precision and generalization. Extensive experiments on the UnAV-100 and LLP datasets show PreFM significantly outperforms state-of-the-art methods by a large margin with significantly fewer parameters, offering an insightful approach for real-time multimodal video understanding.

Humans naturally understand moments in a video by integrating visual and auditory cues. For example, localizing a scene in the video like “A scientist passionately speaks on wildlife conservation as dramatic orchestral music plays, with the audience nodding and applauding” requires simultaneous processing of visual, audio, and speech signals. However, existing models often struggle to effectively fuse and interpret audio information, limiting their capacity for comprehensive video temporal understanding. To address this, we present TriSense, a triple-modality large language model designed for holistic video temporal understanding through the integration of visual, audio, and speech modalities. Central to TriSense is a Query-Based Connector that adaptively reweights modality contributions based on the input query, enabling robust performance under modality dropout and allowing flexible combinations of available inputs. To support TriSense's multimodal capabilities, we introduce TriSense-2M, a high-quality dataset of over 2 million curated samples generated via an automated pipeline powered by fine-tuned LLMs. TriSense-2M includes long-form videos and diverse modality combinations, facilitating broad generalization. Extensive experiments across multiple benchmarks demonstrate the effectiveness of TriSense and its potential to advance multimodal video analysis.

Estimating motion primitives from video (e.g., optical flow and occlusion) is a critically important computer vision problem with many downstream applications, including controllable video generation and robotics. Current solutions are primarily supervised on synthetic data or require tuning of situation-specific heuristics, which inherently limits these models' capabilities in real-world contexts. A natural solution to transcend these limitations would be to deploy large-scale, self-supervised video models, which can be trained scalably on unrestricted real-world video datasets. However, despite recent progress, motion-primitive extraction from large pretrained video models remains relatively underexplored. In this work, we describe Opt-CWM, a self-supervised flow and occlusion estimation technique from a pretrained video prediction model. Opt-CWM uses ``counterfactual probes'' to extract motion information from a base video model in a zero-shot fashion. The key problem we solve is optimizing the quality of these probes, using a combination of an efficient parameterization of the space counterfactual probes, together with a novel generic sparse-prediction principle for learning the probe-generation parameters in a self-supervised fashion. Opt-CWM achieves state-of-the-art performance for motion estimation on real-world videos while requiring no labeled data.

The neural activity in the visual processing is influenced by both external stimuli and internal brain states. Ideally, a neural predictive model should account for both of them. Currently, there are no dynamic encoding models that explicitly model a latent state and the entire neuronal response distribution. We address this gap by proposing a probabilistic model that predicts the joint distribution of the neuronal responses from video stimuli and stimulus-independent latent factors. After training and testing our model on mouse V1 neuronal responses, we find that it outperforms video-only models in terms of log-likelihood and achieves improvements in likelihood and correlation when conditioned on responses from other neurons. Furthermore, we find that the learned latent factors strongly correlate with mouse behavior and that they exhibit patterns related to the neurons’ position on the visual cortex, although the model was trained without behavior and cortical coordinates. Our findings demonstrate that unsupervised learning of latent factors from population responses can reveal biologically meaningful structure that bridges sensory processing and behavior, without requiring explicit behavioral annotations during training. The code is attached to the submission.

Humans can efficiently extract knowledge and learn skills from the videos within only a few trials and errors. However, it poses a big challenge to replicate this learning process for autonomous agents, due to the complexity of visual input, the absence of action or reward signals, and the limitations of interaction steps. In this paper, we propose a novel, unsupervised, and sample-efficient framework to achieve imitation learning from videos (ILV), named Behavior Cloning from Videos via Latent Representations (BCV-LR). BCV-LR extracts action-related latent features from high-dimensional video inputs through self-supervised tasks, and then leverages a dynamics-based unsupervised objective to predict latent actions between consecutive frames. The pre-trained latent actions are fine-tuned and efficiently aligned to the real action space online (with collected interactions) for policy behavior cloning. The cloned policy in turn enriches the agent experience for further latent action finetuning, resulting in an iterative policy improvement that is highly sample-efficient. We conduct extensive experiments on a set of challenging visual tasks, including both discrete control and continuous control. BCV-LR enables effective (even expert-level on some tasks) policy performance with only a few interactions, surpassing state-of-the-art ILV baselines and reinforcement learning methods (provided with environmental rewards) in terms of sample efficiency across 24/28 tasks. To the best of our knowledge, this work for the first time demonstrates that videos can support extremely sample-efficient visual policy learning, without the need to access any other expert supervision.