A challenge in advancing Visual-Language Models (VLMs) is determining whether their failures on abstract reasoning tasks, such as Bongard problems, stem from flawed perception or faulty top-down reasoning. To disentangle these factors, we introduce a diagnostic framework centered on the Linear Separability Ceiling (LSC), the performance achievable by a linear classifier on a VLM's raw visual embeddings. Applying this framework to state-of-the-art VLMs, we uncover a pervasive ''alignment gap'', where most models fail to generatively outperform the linear separability of their representations. We find that the few models surpassing this ceiling do so via two mechanisms: by further refining visual representations into a more linearly separable format or by executing non-linear decision logic. We demonstrate that this bottleneck is not a fundamental limitation but a solvable visual alignment issue. Our method augments standard next-token prediction with a contrastive objective to restructure the visual manifold into a more one-dimensionally linear geometry, improving image-to-image comparison and enabling models to significantly surpass the LSC on abstract compositional reasoning tasks.

Recent advancements in Multimodal Large Language Models (MLLMs) have achieved remarkable success in vision-language tasks, yet the quadratic computational complexity arising from the vast number of visual tokens incurs significant memory and latency bottlenecks. While visual token reduction (VTR) strategies have been explored to mitigate this burden, existing methods overlook the positional and attentional consistency between the full and reduced sequences, resulting in a distorted representation. To this end, we propose RESTORE, a novel VTR framework that rectifies the positional and attentional distortions while maintaining efficiency. Specifically, we present a simple yet effective calibration method that restores lost visual attention by augmenting attention weights based on relative distances. We also introduce a distinctive anchor selection for token merging to mitigate information loss during feature averaging. Experimental results on multiple benchmarks demonstrate that our method consistently improves the accuracy of various reduction methods, achieving state-of-the-art performance while maintaining computational efficiency. Project page is available at https://cvlab.yonsei.ac.kr/projects/RESTORE

We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-transformers architecture. By supporting highly flexible input-output configurations, Cosmos 3 seamlessly unifies critical modalities for Physical AI -- effectively subsuming vision-language models, video generators, world simulators, and world-action models into a single framework. Our evaluation demonstrates that Cosmos 3 establishes a new state-of-the-art across a diverse suite of understanding and generation tasks, demonstrating omnimodal world models as scalable, general-purpose backbones for embodied agents. Our post-trained Cosmos 3 models were ranked as the best open-source Text-to-Image and Image-to-Video models by Artificial Analysis, and the best policy model by RoboArena at the time the technical report was written. To accelerate open research and deployment in Physical AI, we make our code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under the Linux Foundation's OpenMDW-1.1 License at https://github.com/nvidia/cosmos and https://huggingface.co/collections/nvidia/cosmos3. The project website is available at https://research.nvidia.com/labs/cosmos-lab/cosmos3.

Vision-language models (VLMs), like human learners, are frequently exposed to new visual concepts, but how they map novel visual references to language after exposure remains largely underexplored, particularly when those references contradict prior knowledge from pre-training. To study this, we present the Novel Visual References Dataset (NVRD): 19,176 images spanning 90 visual concepts across different levels of visual novelty, each with up to 20 increasingly perturbed versions of the original object to probe generalization. Unlike prior work on visual augmentations of familiar concepts, NVRD comprises entirely novel, open-ended stimuli constructed from scratch, mirroring how humans encounter genuinely new concepts. We evaluate 3 open- and 2 closed-source models alongside 2,400 human judgments for direct human-model comparison, and find that (i) models struggle to acquire novel concepts in-context when they contradict prior knowledge, and (ii) while models and humans show correlated sensitivity to visual perturbations, models significantly overgeneralize, extending learned labels to stimuli that humans reject. We contribute NVRD as a corpus and benchmark for research on visual concept learning in both humans and machines.

Although Multimodal Large Language Models (MLLMs) demonstrate strong omni-modal perception, their ability to forecast future events from audio-visual cues remains largely unexplored, as existing benchmarks focus mainly on retrospective understanding. To bridge this gap, we introduce FutureOmni, the first benchmark designed to evaluate omni-modal future forecasting from audio-visual environments. The evaluated models are required to perform cross-modal causal and temporal reasoning, as well as effectively leverage internal knowledge to predict future events. FutureOmni is constructed via a scalable LLM-assisted, human-in-the-loop pipeline and contains 919 videos and 1,034 multiple-choice QA pairs across 8 primary domains. Evaluations on 13 omni-modal and 7 video-only models show that current systems struggle with audio-visual future prediction, particularly in speech-heavy scenarios, with the best accuracy of 64.8% achieved by Gemini 3 Flash. To mitigate this limitation, we curate a 7K-sample instruction-tuning dataset and propose an Omni-Modal Future Forecasting (OFF) training strategy. Evaluations on FutureOmni and popular audio-visual and video-only benchmarks demonstrate that OFF enhances future forecasting and generalization. We publicly release all code (https://github.com/OpenMOSS/FutureOmni) and datasets (https://huggingface.co/datasets/OpenMOSS-Team/FutureOmni).

End-to-end autonomous driving models are typically trained on multi-city datasets using supervised ImageNet-pretrained backbones, yet their ability to generalize to unseen cities remains largely unexamined. When training and evaluation data are geographically mixed, models may implicitly rely on city-specific cues, masking failure modes that would occur under real-world domain shifts when generalizing to new locations. In this work, we formulate zero-shot cross-city transfer as a controlled representation-level stress test for end-to-end autonomous driving and ask how visual pretraining affects transfer behavior under geographic domain shift. We conduct a comprehensive study by integrating self-supervised backbones I-JEPA, DINOv2, and MAE into planning frameworks. We evaluate performance under strict geographic splits on nuScenes in the open-loop setting and on NAVSIM in the closed-loop evaluation protocol. Our experiments reveal a substantial generalization gap when transferring models across cities with different road topologies, traffic conventions, and visual environments. In open-loop evaluation, a supervised backbone exhibits severe degradation when transferring between cities, yet some domain-specific self-supervised methods can substantially reduce both displacement and collision degradation. In closed-loop evaluation, self-supervised pretraining improves average out-of-distribution PDMS in several single-city training settings. Our results provide empirical evidence that representation learning influences the robustness of cross-city planning and motivate zero-shot geographic transfer as an important stress test for evaluating end-to-end autonomous driving systems.

We introduce Qwen-RobotWorld, a language-conditioned video world model for embodied intelligence. With natural language as a unified action interface, it predicts physically grounded future visual trajectories from current observations across robotic manipulation, autonomous driving, indoor navigation, and human-to-robot transfer. This unified formulation provides three promising application directions: synthetic data generation for policy training augmentation, scalable virtual environments for policy evaluation, and language-guided planning signals for downstream robot control. This is achieved through a three-part design: a) Double-Stream MMDiT with MLLM Action Encoding, where a 60-layer double-stream diffusion transformer couples frozen Qwen2.5-VL semantics with video-VAE latents through layer-wise joint attention; b) Embodied World Knowledge (EWK), an 8.6M video-text corpus (200M+ frames) with action-language mapping over 20+ embodiments and 500+ action categories; and c) General+Expert Progressive Curriculum, a two-stage training strategy that first learns general visual priors and then injects embodied specialization under a shared language interface. Extensive results show strong competitiveness: ranks 1st overall on EWMBench and DreamGen Bench, outperforms all open-source models on WorldModelBench and PBench. Additional zero-shot analyses on RoboTwin-IF benchmark further support robust generalization and multi-view consistency.

Foundation models in language and multimodality achieve strong generalization by aligning heterogeneous data under a unified formulation and training at scale. In this report, we investigate whether this scaling recipe can be applied to robotic manipulation to achieve genuine generalization. This is challenging because, unlike text, manipulation data is heterogeneous by nature, expensive to collect, and narrow in diversity, making alignment and scale simultaneously difficult. We present Qwen-RobotManip, a generalizable Vision-Language-Action foundation model built on Qwen-VL. Qwen-RobotManip introduces a unified alignment framework across the representation, motion, and behavioral dimensions of manipulation, making large-scale multi-source training coherent rather than conflicting. This alignment capability in turn enables Qwen-RobotManip to absorb manipulation data at a scale that prior training regimes could not sustain. A human-to-robot synthesis pipeline converts egocentric hand demonstrations into robot trajectories across 15 platforms, and a rigorous curation pipeline harmonizes heterogeneous datasets. Using only open-source datasets and human videos without proprietary data collection, Qwen-RobotManip constructs a ~38,100-hour pretraining corpus and exhibits emergent generalization capabilities, including zero-shot instruction following, robustness to perturbations, reactive error recovery, and cross-embodiment transfer. We find that standard benchmarks fail to capture pretraining quality and instead adopt OOD settings including RoboCasa365, LIBERO-Plus, EBench, RoboTwin-Clean2Rand, RoboTwin-IF, and RoboTwin-XE. Qwen-RobotManip substantially outperforms prior state-of-the-art models, including $π$0.5, across all OOD settings, ranks 1st in RoboChallenge with a 20% relative improvement, and is validated on real-robot platforms including AgileX ALOHA, Franka, UR, and ARX.

Artificial intelligence provides a practical framework for crop damage assessment from imagery data, supporting early decision-making in agricultural management. In peach orchards, climate change increases abiotic stress and biotic pressures, including pests and diseases, which often produce visually similar foliar symptoms. This overlap makes manual diagnosis difficult, especially across multiple fields with varying environmental conditions, highlighting the need for automated models with strong generalization ability. We propose an image-based classification approach for peach leaf damage detection. A benchmark dataset was created through manual annotation of publicly available images, consisting of 1,366 peach leaves across six damage categories. Several deep learning architectures were evaluated. EfficientNet models achieved the best results, with EfficientNetB0 reaching 92.9 percent accuracy, EfficientNetB3 achieving 91.5 percent, and EfficientNetB5 showing the strongest performance on minority classes. DenseNet121 reached 92.6 percent accuracy. The integration of the Convolutional Block Attention Module (CBAM) improved performance in several backbones, particularly EfficientNetB5 and InceptionV3, while showing limited or negative impact in others. The CBAM-enhanced EfficientNetB5 achieved the best overall accuracy of 93.3 percent. To evaluate robustness under realistic conditions, a local dataset of 180 images across four classes was collected, and transfer learning strategies were applied to address domain shift. Three fine-tuning strategies were tested. EfficientNetB3 combined with CBAM achieved the best performance in the local domain, reaching a 93 percent macro F1-score after transfer. Overall, attention-based models showed improved robustness for minority classes and better generalization across different field conditions.

Visual Place Recognition (VPR) is a key component for localisation in Global Navigation Satellite System (GNSS)-denied environments, but its performance critically depends on selecting an image matching threshold (operating point) that balances precision and recall. Thresholds are typically hand-tuned offline for a specific environment and fixed during deployment, leading to degraded performance under environmental change. We propose a method that automatically selects the operating point of a VPR system to maximise recall at 100% precision. The method uses a small calibration traversal with known correspondences and transfers thresholds to deployment via quantile normalisation of similarity score distributions. This quantile transfer ensures that thresholds remain stable across calibration sizes and query subsets. Experiments with seven state-of-the-art VPR techniques across five benchmark datasets demonstrate that our proposed approach consistently outperforms existing baselines, enabling the underlying VPR technique to operate at 100% precision in approximately twice as many deployment scenarios (median improvement), while retrieving up to 29% more correct matches at that precision. The method eliminates manual tuning by adapting to new environments and generalising across operating conditions. Our code is available at https://github.com/DhyeyR-007/Quantile-Transfer-for-Reliable-VPR.

In Remote Sensing (RS), Parameter-Efficient Fine-Tuning (PEFT) has emerged as a key approach to activate the generalizable representation ability of foundation models for downstream tasks. However, existing specialized PEFT methods often fail when applied to large-scale Earth observation tasks, as they are unable to fully handle the multifaceted and unpredictable domain gaps (e.g., spatial, semantic, and frequency shifts) inherent in RS data. To overcome this, we propose CrossEarth-Gate, which introduces two primary contributions. First, we establish a comprehensive RS module toolbox to address multifaceted domain gaps, comprising spatial, semantic, and frequency modules. Second, we develop a Fisher-guided adaptive selection mechanism that operates on this toolbox. This selection is guided by Fisher Information to quantify each module's importance by measuring its contribution to the task-specific gradient flow. It dynamically activates only the most critical modules at the appropriate layers, guiding the gradient flow to maximize adaptation effectiveness and efficiency. Comprehensive experiments validate the efficacy and generalizability of our method, where CrossEarth-Gate achieves state-of-the-art performance on 16 out of 18 cross-domain benchmarks for RS semantic segmentation.

Unsupervised object-centric learning (OCL) decomposes visual scenes into distinct entities. Slot attention is a popular approach that represents individual objects as latent vectors, called slots. Current methods obtain these slot representations solely from the last layer of a pre-trained vision transformer (ViT), ignoring valuable, semantically rich information encoded across the other layers. To better utilize this latent semantic information, we introduce MUFASA, a lightweight plug-and-play framework for slot-attention-based approaches to unsupervised object segmentation. Our model computes slot attention across multiple feature layers of the ViT encoder, fully leveraging their semantic richness. We propose a fusion strategy to aggregate slots obtained on multiple layers into a unified object-centric representation. Integrating MUFASA into existing OCL methods improves their segmentation results across multiple datasets, setting a new state of the art while simultaneously improving training convergence with only minor inference overhead.

Multimodal semantic segmentation with heterogeneous sensors must reconcile complementary information across modalities that differ in spatial resolution and channel dimensionality. In particular, high-resolution RGB imaging provides detailed spatial structure but often fails to distinguish visually similar materials, whereas hyperspectral imaging (HSI) provides discriminative spectral signatures but at lower spatial resolution. We present Bidirectional Cross-Attention Fusion (BCAF), which aligns high-resolution RGB with low-resolution HSI at their native grids via localized, bidirectional cross-attention, avoiding pre-upsampling or early spectral collapse. BCAF uses two independent backbones: a standard Swin Transformer for RGB and an HSI-adapted Swin backbone that preserves spectral structure through 3D tokenization with spectral self-attention. Although our evaluation targets RGB-HSI fusion, BCAF is modality-agnostic and applies to co-registered RGB with lower-resolution, high-channel auxiliary sensors. On the benchmark SpectralWaste dataset, BCAF delivers strong performance, achieving 75.4% at 55 images/s. We further evaluate a novel industrial dataset: K3I-Cycling (first RGB subset already released on Fordatis). On this dataset, BCAF reaches 62.3% mIoU for material segmentation (paper, metal, plastic, etc.) and 66.2% mIoU for plastic-type segmentation (PET, PP, HDPE, LDPE, PS, etc.). These results show that preserving native-grid spatial detail and spectral structure improves multimodal segmentation under real-time constraints. Code and model checkpoints are publicly available at https://github.com/jonasvilhofunk/BCAF_2026.

Multimodal change detection (MMCD) identifies changed areas in multimodal remote sensing data, demonstrating significant application value in land use monitoring and urban sustainable development. However, literature MMCD approaches exhibit limitations in both cross-modal interaction and exploiting modality-specific characteristics. This leads to insufficient modeling of fine-grained change information, thus hindering the precise detection of semantic changes. To address these problems, we propose STSF-Net, a framework designed for MMCD between optical and SAR images. STSF-Net jointly models modality-specific and spatio-temporal common features to enhance change representations. Specifically, modality-specific features are exploited to capture genuine semantic change signals, while spatio-temporal common features are embedded to suppress pseudo-changes caused by differences in imaging mechanisms. Furthermore, we introduce an optical and SAR feature fusion strategy that adaptively adjusts multimodal feature importance based on semantic priors obtained from visual foundation models. Finally, we introduce the novel Delta-SN6 dataset, the first openly-accessible multiclass MMCD benchmark consisting of very-high-resolution fully polarimetric SAR and optical images. Experimental results on Delta-SN6, BRIGHT, and Wuhan datasets demonstrate that our method outperforms the state-of-the-art by 3.21%, 0.87%, and 1.32% in mIoU, respectively.

We present SegmentAnyTreeV2, a sensor- and platform-agnostic framework for semantic and instance segmentation of forest point clouds. The model combines a serialization-based Point Transformer v3 backbone with a lightweight semantic head and a tree-focused cross-attention mask decoder. Semantic predictions restrict instance decoding to tree-class voxels, while instance-aware query initialization, one-to-many seed supervision, and asymmetric mask scoring improve separation in dense and structurally complex stands. We further introduce FOR-instance v3, an expanded benchmark comprising 427 scenes and 26,496 annotated trees across diverse biomes, forest structures, and LiDAR platforms. On the FOR-instanceV2 test split, SegmentAnyTreeV2 achieves 90.5% precision, 80.2% recall, 85.0% F1, 90.7% coverage, and 87.6% semantic mIoU, outperforming previous learning-based methods in both instance detection and mask completeness. Zero-shot evaluation on independent sites further demonstrates strong cross-domain generalization.

Disciplines such as business process management and process mining aid organizations by discovering insights about processes on the basis of recorded event data. However, an obstacle to process analysis is data multi-modality: for instance, data in video form are not directly interpretable as events. Existing approaches rely on a dictionary of activity label as input, cannot provide frame-by-frame labeling explanations, or rely on superseded computer vision techniques. In this work, we present SnapLog, an approach to extract event data from videos by converting frames to feature vectors using image embeddings and performing temporal segmentation through frame-wise similarity matrices. A generalized few-shot classification is then used to assign labels to the video segments, yielding labeled, timestamped sub-sequences of frames that are interpretable as events. Conventional process mining techniques can be used to analyze the resulting data. We show that our approach produces logs that accurately reflect the process in the videos.

While highlight detection for long-form videos is of great practical importance, most existing methods remain limited to short-form content, largely due to the absence of a suitable benchmark. To bridge this gap, we introduce SVHighlights, to the best of our knowledge, the first benchmark for highlight detection in extremely long sports videos, each exceeding one hour in duration, across multiple sports categories. SVHighlights is constructed from pairs of full-length sports videos and their corresponding official highlight videos using a dataset generation pipeline, enabling scalable label generation without conventional per-clip saliency annotation. The benchmark comprises 320 videos with an average duration of 2.00 hours and a total of 640.18 hours, substantially exceeding previous datasets. Existing methods also face fundamental challenges on long videos: models trained on short clips fail to generalize to hour-long content, and their clip-level scoring lacks the broader context needed to identify highlights. To address this and provide a strong baseline, we present TF-SELECTOR, a training-free segment-based approach that divides each video into context-aware segments by merging adjacent shots sharing the same semantic content, and predicts segment-level saliency scores using a large language model with multimodal inputs including visual captions, transcripts, and audio volume. Experiments demonstrate that TF-SELECTOR achieves superior performance across most metrics compared to Video Temporal Grounding (VTG)-tuned baselines, with improvements of +2.50 in HIT@1, +4.04 in HIT@K, and +2.95 in IoU. These results establish SVHighlights as a challenging testbed for long-form highlight detection and demonstrate that a simple segment-based strategy can effectively scale to hour-long videos.

While video segmentation has advanced rapidly on short clips and closed-set benchmarks, open-world video segmentation remains largely unexplored. The challenge is twofold: (1) existing methods are not designed to support object discovery and identity maintenance in long videos of dynamic ego-motion, and (2) existing evaluation protocols rely on a rigid 1:1 matching that unfairly penalizes semantically valid predictions with mismatched granularity. To address both gaps, we introduce Savvy, a practical and strong system for zero-shot open-world long-horizon video segmentation. Savvy combines hierarchical mask discovery, deferred admission, and track consolidation to support persistent object discovery, safe track promotion, and stable long-range identity maintenance. We further propose OGA, a granularity-aware evaluation suite for open-world video segmentation. Built on a Granularity-Agnostic (GA) matching protocol, OGA relaxes conventional 1:1 matching to an n:1 mapping, but still enforces temporal rigor by detecting support discontinuities through sever points and scoring each reference object through its dominant coherent fragment. This prevents fragmented or flickering support from being over-rewarded while enabling GA-adapted metrics and structural diagnostics: identity persistence (IP), and identity concentration (IC). On VIPSeg, we show that standard 1:1 evaluation substantially underestimates open-world methods, whereas GA evaluation recovers much of their suppressed performance. On the more realistic long-horizon benchmarks: ScanNet and HM3D, Savvy consistently outperforms strong baselines across both classical and proposed metrics, including STQ, VPQ$_\infty$, IP and IC. Together, these results establish a practical benchmark and a strong baseline for open-world long-horizon video segmentation.

Controllable image-to-video (I2V) generation transforms a reference image into a coherent video guided by user-specified control signals. While precise control over camera motion, object motion, and lighting is essential for high-fidelity creation, existing methods often treat these factors independently. This overlooks the physical coupling among viewpoint, geometry, and illumination in dynamic scenes, leading to visual inconsistencies such as mismatched shadows and perspective drift under simultaneous changes. We present VidCRAFT3, a unified and flexible I2V framework that explicitly models cross-factor interactions among geometry, motion, and illumination, enabling both independent and joint control over camera motion, object motion, and lighting direction. Image2Cloud provides explicit 3D geometric priors for accurate camera motion control. ObjMotionNet encodes sparse object trajectories into multi-scale motion features to guide realistic object motion. A Spatial Triple-Attention Transformer integrates lighting direction through lighting cross-attention for consistent relighting. To address the scarcity of jointly annotated data, we construct the VideoLightingDirection (VLD) dataset with accurate per-frame lighting direction annotations, and introduce a three-stage progressive training strategy that enables robust learning without fully joint annotations. Extensive experiments demonstrate that VidCRAFT3 achieves state-of-the-art performance in control precision and visual coherence across diverse scenarios.

Video generation models have advanced significantly through the latent diffusion transformers trained with rectified flow techniques. Yet these models still struggle with geometric inconsistencies, unstable motion, and visual artifacts that break the illusion of realistic 3D scenes. 3D-consistent video generation could significantly impact numerous downstream applications in generation and reconstruction tasks. We explore how epipolar geometry constraints improve modern video diffusion models. Despite using massive training data, these models fail to capture fundamental geometric principles. We align diffusion models using pairwise epipolar geometry constraints via preference-based optimization, directly addressing unstable trajectories and geometric artifacts through mathematically principled geometric enforcement. Our approach efficiently enforces geometric principles without requiring end-to-end differentiability. Evaluation demonstrates that classical geometric constraints provide more stable optimization signals than modern learned metrics. Training on static scenes with dynamic cameras ensures metric quality while the model generalizes to various dynamic scenes. By bridging data-driven learning with classical computer vision, we reduce epipolar error by 31% and improve human-rated consistency from 54% to 72% without compromising visual quality.

Conditional image editing aims to modify a source image according to textual prompts and optional reference guidance. Such editing is crucial in scenarios requiring strict structural control (i.e., anomaly insertion in driving scenes and complex human pose transformation). Despite recent advances in large-scale editing models (i.e., Seedream, Nano Banana, etc), most approaches rely on single-step generation. This paradigm often lacks explicit quality control, may introduce excessive deviation from the original image, and frequently produces structural artifacts or environment-inconsistent modifications, typically requiring manual prompt tuning to achieve acceptable results. We propose \textbf{CAMEO}, a structured multi-agent framework that reformulates conditional editing as a quality-aware, feedback-driven process rather than a one-shot generation task. CAMEO decomposes editing into coordinated stages of planning, structured prompting, hypothesis generation, and adaptive reference grounding, where external guidance is invoked only when task complexity requires it. To overcome the lack of intrinsic quality control in existing methods, evaluation is embedded directly within the editing loop. Intermediate results are iteratively refined through structured feedback, forming a closed-loop process that progressively corrects structural and contextual inconsistencies. We evaluate CAMEO on anomaly insertion and human pose switching tasks. Across multiple strong editing backbones and independent evaluation models, CAMEO consistently achieves 20\% more win rate on average compared to multiple state-of-the-art models, demonstrating improved robustness, controllability, and structural reliability in conditional image editing.

Visual Autoregressive (VAR) models have recently demonstrated impressive image generation quality while maintaining low latency. However, they suffer from severe KV-cache memory constraints, often requiring gigabytes of memory per generated image. We introduce HeatKV, a novel compression method that adapts cache allocation in each head based on its attention to previously generated scales. Using a small offline calibration set, the attention heads are ranked according to their attention scores over prior scales. Based on this ranking, we construct a static pruning schedule tailored to a given memory budget. Applied to the Infinity-2B model, HeatKV achieves $2 \times$ higher compression ratio in memory allocation for KV cache compared to existing methods, while maintaining similar or better image fidelity, prompt alignment and human perception score. Our method achieves a new state-of-the-art (SOTA) for VAR model KV-cache compression, showcasing the effectiveness of fine-grained, head-specific cache allocation. Code and calibration script available at https://github.com/arm-research/heatkv.

Human-centric video customization, particularly at the garment level, has shown significant commercial value. However, existing approaches cannot support low-latency and interactive garment control, which is crucial for applications such as e-commerce and content creation. This paper studies how to achieve interactive multi-garment video customization while preserving motion coherence using only single-garment video data. We present FashionChameleon, a real-time and interactive framework for human-garment customization in autoregressive video generation, where users can interactively switch garment during generation. FashionChameleon consists of three key techniques: (i) Instead of training on multi-garment video data, we train a Teacher Model with In-Context Learning on a single reference-garment pair. By retaining the image-to-video training paradigm while enforcing a mismatch between the reference and garment image, the model is encouraged to implicitly preserve coherence during single-garment switching. (ii) To achieve consistency and efficiency during generation, we introduce Streaming Distillation with In-Context Learning, which fine-tunes the model with in-context teacher forcing and improves extrapolation consistency via gradient-reweighted distribution matching distillation. (iii) To extend the model for interactive multi-garment video customization, we propose Training-Free KV Cache Rescheduling, which includes garment KV refresh, historical KV withdraw, and reference KV disentangle to achieve garment switching while preserving motion coherence. Our FashionChameleon uniquely supports interactive customization and consistent long-video extrapolation, while achieving real-time generation at 23.8 FPS on a single GPU, 30-180$\times$ faster than existing baselines.

Autoregressive long video generation often adopts bounded-memory streaming for efficiency, typically combining local windows for short-term continuity with static early-frame sinks as long-range anchors. However, this fixed allocation keeps early frames cached even when the current visual state has substantially diverged from them, while discarding potentially more relevant intermediate history. As a result, the retained long-range context may become less adaptive and bias generation toward outdated cues; in severe cases, RoPE-induced phase re-alignment can homogenize inter-head attention and cause sink collapse, where content regresses toward sink frames. We propose DySink, a retrieval-based framework that maintains a compact memory bank and selects visually relevant historical frames as dynamic frame sinks. DySink couples adaptive retrieval with a sink anomaly gate, which detects excessive inter-head consensus over retrieved context and suppresses collapse-prone context. Experiments on minute-long videos show that DySink consistently improves temporal quality over strong baselines while also achieving higher dynamic degree, enabling coherent and more natural long-horizon visual evolution. The code and model weights are released at https://github.com/yebo0216best/DySink.

Video Virtual Try-On (VVT) aims to seamlessly replace a garment on a person in a video with a new one. While existing methods have made significant strides in maintaining temporal consistency, they are predominantly confined to non-interactive scenarios where models merely showcase garments. This limitation overlooks a crucial aspect of real-world apparel presentation: active human-garment interaction. To bridge this gap, we introduce and formalize a new challenging task: Interactive Video Virtual Try-On (Interactive VVT), where subjects in the video actively engage with their clothing. This task introduces unique challenges beyond simple texture preservation, including: (1) resolving the semantic ambiguity of interactions from standard pose information, and (2) learning complex garment deformations from video where interactive moments are sparse and brief. To address these challenges, we propose iTryOn, a novel framework built upon a large-scale video diffusion Transformer. iTryOn pioneers a multi-level interaction injection mechanism to guide the generation of complex dynamics. At the spatial level, we introduce a garment-agnostic 3D hand prior to provide fine-grained guidance for precise hand-garment contact, effectively resolving spatial ambiguity. At the semantic level, iTryOn leverages global captions for overall context and time-stamped action captions for localized interactions, synchronized via our novel Action-aware Rotational Position Embedding (A-RoPE). Extensive experiments demonstrate that iTryOn not only achieves state-of-the-art performance on traditional VVT benchmarks but also establishes a commanding lead in the new interactive setting, marking a significant step towards more dynamic and controllable virtual try-on experiences.

Image-to-Video diffusion models leverage input images to generate visually stunning content, yet frequently produce motion that violates physical laws. We reveal a surprising finding: a 2-step generation often exhibits better physical consistency than a 50-step output from the same model. Through spectral analysis, we trace this to phase erosion during denoising; the phase degrades significantly (dropping by $\approx 18\%$ from step 2 to step 50), whereas the magnitude remains relatively stable. Building on this insight, we propose PhaseLock, a training-free framework that preserves the valid motion priors from few-step inference throughout the denoising trajectory. Rather than relying on full-step inference for physical consistency, PhaseLock extracts a motion prior from just 2 steps and enforces it onto high-fidelity generation via Latent Delta Guidance. Our approach effectively mitigates phase degradation, improving physical consistency by an average of 6.2 points across diverse models while largely maintaining visual fidelity, with negligible overhead ($1.06\times$ time, $1.02\times$ memory) and reduced reliance on expensive external guidance methods ($\sim5\times$ time). Project Page: https://dnwjddl.github.io/phaselock

We present MoVerse, a real-time video world model that creates an interactively navigable scene from a single narrow-field-of-view image. This setting is challenging because the input observes only a small fraction of the environment, while interactive roaming requires a complete surrounding world, persistent geometry, controllable camera motion, and temporally coherent high-fidelity observations. MoVerse addresses this problem by separating world construction from observation rendering. It first expands the input into a gravity-aligned 360$^\circ$ panorama with topology-aware diffusion, closing the missing field of view before 3D reasoning. It then lifts the panorama into a persistent 3D Gaussian scaffold using panoramic geometry-aware residual prediction, yielding a dense and directly renderable spatial memory. Finally, a Gaussian-conditioned video renderer translates scaffold renderings along user-specified camera trajectories into photorealistic video. To make this renderer practical for interaction, we train a bidirectional diffusion teacher for high-quality conditional rendering and distill it into a causal autoregressive student for bounded-latency streaming. This design combines the controllability and long-range consistency of explicit 3D representations with the perceptual quality of generative video models. MoVerse supports real-time scene roaming at 8~FPS on a single NVIDIA RTX~4090 GPU, demonstrating a practical path toward single-image world creation with interactive video output.

Cinematic video depicts multiple subjects acting or interacting at specific moments, captured with deliberate camera movement, and stitched together by shot transitions. Together, these elements demand a level of fine-grained control beyond current text-to-video models. Existing work addresses each axis in isolation: multi-subject personalization, temporal control, multi-shot synthesis, or camera control; no prior framework jointly integrates all four. We present CineOrchestra, a unified video diffusion model that controls subjects, events, cameras, and shot transitions simultaneously. Our key insight is that these heterogeneous cinematic elements share a fundamental structure: each is an entity acting over a specific temporal interval, which can therefore all be expressed through one shared structure of entity-centric conditioning primitives, augmented with reference images for visual entities. This formulation reduces the architectural challenge to a single positional encoding problem, which we solve with two parameter-free coordinated rotary embeddings: (a) an interval-sampled temporal RoPE that yields consistent attention behavior across events of dramatically varying duration, and (b) a 2D entity-temporal cross-attention RoPE that disambiguates per-entity conditions and routes each to its corresponding spatiotemporal region. On two new benchmarks, CineOrchestra outperforms six per-axis specialists on dense caption following and shot-transition timing, with consistent gains in a pairwise user study and component ablations. Project page: https://snap-research.github.io/CineOrchestra

Conventional production workflow of high-precision mesh assets necessitates a cumbersome and laborious process of manual sculpting by specialized 3D artists/modelers. The recent years have witnessed remarkable advances in AI-empowered 3D content creation for generating plausible structures and intricate appearances from images or text prompts. However, synthesizing realistic surface details still poses great challenges, and enhancing the geometry fidelity of existing lower-quality 3D meshes (instead of image/text-to-3D generation) remains an open problem. In this paper, we introduce SuperCarver, a 3D geometry super-resolution pipeline for supplementing texture-consistent surface details onto a given coarse mesh. We start by rendering the original textured mesh into the image domain from multiple viewpoints. To achieve detail boosting, we construct a deterministic prior-guided normal diffusion model, which is fine-tuned on a carefully curated dataset of paired detail-lacking and detail-rich normal map renderings. To update mesh surfaces from potentially imperfect normal map predictions, we design a noise-resistant inverse rendering scheme through deformable distance field. Experiments demonstrate that our SuperCarver is capable of generating realistic and expressive surface details depicted by the actual texture appearance, making it a powerful tool to both upgrade historical low-quality 3D assets and reduce the workload of sculpting high-poly meshes.

Point cloud stands as the most widely adopted format for representing 3D shapes and scenes due to its simplicity and geometric fidelity. However, its inherent unordered and irregular nature, exacerbated by sensor noise and occlusions, introduces unique challenges for machine learning based methodologies. To combat these issues, diverse strategies have been developed, including converting to a format that has orderliness, extracting local geometry, and permutation-invariant or self-attention-based processing. In this paper, our focus is directed towards deep learning models for three fundamental tasks in 3D vision: point cloud classification, part segmentation, and semantic segmentation. We begin by formally defining point cloud data, followed by an in-depth discussion on its structural characteristics. Then, we categorize notable works based on their backbone structure and evaluate their performance on popular benchmarks. Beyond empirical comparison, we offer insights into architectural innovations and limitations. We also outline open challenges and promising future directions for 3D point cloud understanding.

With the growing volume of CT examinations, there is an increasing demand for automated tools such as organ segmentation, abnormality detection, and report generation to support radiologists in managing their clinical workload. Multi-label classification of 3D Chest CT scans remains a critical yet challenging problem due to the complex spatial relationships inherent in volumetric data and the wide variability of abnormalities. Existing methods based on 3D convolutional neural networks struggle to capture long-range dependencies, while Vision Transformers often require extensive pre-training on large-scale, domain-specific datasets to perform competitively. In this work, we propose a 2.5D alternative by introducing a new graph-based framework that represents 3D CT volumes as structured graphs, where axial slice triplets serve as nodes processed through spectral graph convolution, enabling the model to reason over inter-slice dependencies while maintaining complexity compatible with clinical deployment. Our method, trained and evaluated on 3 datasets from independent institutions, achieves strong cross-dataset generalization, and shows competitive performance compared to state-of-the-art visual encoders. We further conduct comprehensive ablation studies to evaluate the impact of various aggregation strategies, edge-weighting schemes, and graph connectivity patterns. Additionally, we demonstrate the broader applicability of our approach through transfer experiments on automated radiology report generation and abdominal CT data.

Understanding disease progression is a central clinical challenge with direct implications for early diagnosis and personalized treatment. While recent generative approaches have attempted to model progression, key mismatches remain: disease dynamics are inherently continuous and monotonic, yet latent representations are often scattered, lacking semantic structure, and diffusion-based models disrupt continuity with random denoising process. In this work, we propose to treat the disease dynamic as a velocity field and leverage Flow Matching (FM) to align the temporal evolution of patient data. Unlike prior methods, it captures the intrinsic dynamic of disease, making the progression more interpretable. However, a key challenge remains: in latent space, Auto-Encoders (AEs) do not guarantee alignment across patients or correlation with clinical-severity indicators (e.g., age and disease conditions). To address this, we propose to learn patient-specific latent alignment, which enforces patient trajectories to lie along a specific axis, with magnitude increasing monotonically with disease severity. This leads to a consistent and semantically meaningful latent space. Together, we present $Δ$-LFM, a framework for modeling patient-specific latent progression with flow matching. Across three longitudinal MRI benchmarks, $Δ$-LFM demonstrates strong empirical performance and, more importantly, offers a new framework for interpreting and visualizing disease dynamics.

Weakly supervised segmentation enables model training from plane-level labels. Existing methods often rely on 2D encoders, neglecting the volumetric nature of medical data. We propose TranSamba, a hybrid Transformer-Mamba architecture designed to capture 3D context via cross-plane modeling. TranSamba augments a Vision Transformer backbone with Cross-Plane Mamba blocks, leveraging linear-time modeling for efficient information exchange across neighboring planes. This exchange improves in-plane self-attention and subsequent attention maps for object localization. TranSamba maintains linear time complexity and constant space complexity with respect to the input volume depth. Extensive experiments on three datasets covering diverse modalities and pathologies show that TranSamba achieves state-of-the-art performance, demonstrating the generalizable efficacy of cross-plane modeling. Code is available at: https://github.com/YihengLyu/TranSamba.

Deep learning-based CT segmentation systems often achieve high accuracy on clean benchmark images, but their performance may degrade under heterogeneous clinical imaging conditions such as noise, resolution loss, contrast variation, intensity shift, and artifacts. This instability can limit reliable deployment in real-world medical imaging workflows. We propose Robustness via Augmented Multi-corruption Pipeline (RAMP), a robustness-oriented augmentation framework for CT segmentation. RAMP combines anatomically constrained spatial perturbations, CT intensity transformations, and stochastic multi-corruption composition to expose models to clinically plausible image degradation during training. Across two CT segmentation evaluation settings, RAMP achieved the strongest corrupted-image performance and the smallest clean-to-corrupted robustness gap. In the five-organ noisy evaluation benchmark, RAMP improved mean corrupted Dice from 0.610 to 0.753 and reduced the robustness gap from 0.264 to 0.064 compared with the nnU-Net baseline. In Abdomen1K, RAMP improved mean corrupted Dice from 0.633 to 0.789 and reduced the robustness gap from 0.290 to 0.070. Although RAMP did not achieve the highest clean-image Dice, it substantially mitigated worst-case segmentation collapse under severe image degradation. These results suggest that multi-corruption augmentation can serve as a practical pre-deployment strategy for improving the reliability of CT segmentation systems in heterogeneous clinical environments.

In-silico trials of medical devices require the generation of virtual populations of anatomies. In cardiovascular applications, virtual anatomy is typically represented as a 3D+t mesh sampled from a generative model. However, most existing mesh generators focus on static anatomy, while sequence models often lack explicit periodicity. To this end, we propose 4D F-MeshLDM, a conditional generative framework comprising a convolutional mesh VAE to encode meshes, a structural latent space that parameterises motion using a truncated Fourier series, and a diffusion prior that learns the latent distribution over Fourier coefficient tokens. By conditioning the diffusion process on clinical covariates via affine modulation, we enable controllable synthesis. Sampling tokens and performing inverse Fourier synthesis yield cycle-consistent latent trajectories, which can be decoded into 3D+t cardiac mesh sequences. Experiments on 5,000 UK Biobank subjects demonstrate that 4D F-MeshLDM outperforms state-of-the-art baselines in anatomical fidelity and achieves near-zero cycle closure error. Furthermore, the generated cohorts accurately preserve clinical functional indices, highlighting the potential of our framework for reliable in-silico cardiac trials.

Multiple Instance Learning (MIL) is a standard paradigm for Whole-Slide Image (WSI) analysis and has achieved strong results in computational pathology. However, most MIL pipelines assume a single "gold" label per slide, which conflicts with clinical practice where substantial inter-pathologist variability is common. Existing multi-annotator learning and label-refinement methods typically estimate global annotator reliability or rely on single-instance assumptions, making them poorly suited to MIL and to localized diagnostic contexts where experts disagree. We propose RaLMPH (Reliability-aware Learning for Multi-Pathologist Harmonization), a MIL-based label reconciliation framework for WSIs annotated by multiple pathologists. RaLMPH introduces a reliability field that jointly models (i) local neighborhood structure in WSI feature space and (ii) expert uncertainty (entropy), enabling per-sample identification of trustworthy reference neighborhoods. Leveraging this field, RaLMPH performs sample-wise local annotator ranking to select reliable opinions per slide and applies an adaptive gating mechanism to fuse labels conditioned on local reliability. Experiments on a clinical WSI dataset with labels from six pathologists, as well as controlled simulated benchmarks, show that RaLMPH consistently outperforms existing approaches. Further analyses clarify how our reliability-aware mechanism improves label reconciliation and downstream MIL performance.

Pathological images are inherently multi-scale, requiring pathologists to integrate evidence from global tissue architecture at low magnification to cellular morphology at higher magnification for accurate diagnosis. While existing pathological datasets for vision-language model (VLM) include various scales, they often lack an explicit cross-scale reasoning objective. This limitation prevents VLMs from capturing essential cross-scale representations and learning evidence-based reasoning. To bridge this gap, we introduce the first cross-scale training and evaluation paradigm that formulates pathology interpretation as multi-magnification reasoning. However, creating such a task reveals a critical challenge: multi-image visual question answering (VQA) is prone to text-only shortcuts, which allow models to guess answers using magnification-dependent artifacts rather than visual evidence. To address this, we propose a leakage-aware curation pipeline that combines adversarial text-only screening with constraint-guided question design. Using this pipeline, we construct Scale-VQA, a high-quality benchmark with 4,685 multiple-choice questions grounded in 2,537 pathology images across multiple magnification levels. Finally, we present ScaleReasoner-R1, a model trained via reinforcement learning to optimize performance on the cross-scale VQA task. ScaleReasoner-R1 achieves state-of-the-art performance on our cross-scale reasoning benchmark and generalizes to SOTA performance on established single-scale benchmarks. Findings suggest that even the limited cross-scale supervision can significantly improve pathological understanding. The code and demos will be open-sourced.

Computed tomography (CT) is a cornerstone imaging modality for non-invasive, high-resolution visualization of internal anatomical structures. However, when the scanned object exceeds the scanner's field of view (FOV), projection data are truncated, resulting in incomplete reconstructions and pronounced artifacts near FOV boundaries. Conventional reconstruction algorithms struggle to recover accurate anatomy from such data, limiting clinical reliability. Deep learning approaches have been explored for FOV extension, with diffusion generative models representing the latest advances in image synthesis. Yet, conventional diffusion models are computationally demanding and slow at inference due to their iterative sampling process. To address these limitations, we propose an efficient CT FOV extension framework based on the image-to-image Schrödinger Bridge (I$^2$SB) diffusion model. Unlike traditional diffusion models that synthesize images from pure Gaussian noise, I$^2$SB learns a direct stochastic mapping between paired limited-FOV and extended-FOV images. This direct correspondence yields a more interpretable and traceable generative process, enhancing anatomical consistency and structural fidelity in reconstructions. I$^2$SB achieves superior quantitative performance, with root-mean-square error (RMSE) values of 49.8 HU on simulated noisy data and 152.0 HU on real data, outperforming state-of-the-art diffusion models such as conditional denoising diffusion probabilistic models (cDDPM) and patch-based diffusion methods. Moreover, its one-step inference enables reconstruction in just 0.19 s per 2D slice, representing over a 700-fold speedup compared to cDDPM (135 s) and surpassing DiffusionGAN (0.58 s), the second fastest. This combination of accuracy and efficiency indicates that I$^2$SB has potential for real-time or clinical deployment.

Digital twins have a major potential to form a significant part of urban management in emergency planning, as they allow more efficient designing of the escape routes, better orientation in exceptional situations, and faster rescue intervention. Nevertheless, creating the twins still remains a largely manual effort, due to a lack of 3D-representations, which are available only in limited amounts for some new buildings. Thus, in this paper we aim to synthesize 3D information from commonly available 2D architectural floor plans. We propose two novel pixel-wise segmentation methods based on the MDA-Unet and MACU-Net architectures with improved skip connections, an attention mechanism, and a training objective together with a reconstruction part of the pipeline, which vectorizes the segmented plans to create a 3D model. The proposed methods are compared with two other state-of-the-art techniques and several benchmark datasets. On the commonly used CubiCasa benchmark dataset, our methods have achieved the mean F1 score of 0.86 over five examined classes, outperforming the other pixel-wise approaches tested. We have also made our code publicly available to support research in the field.

Handwritten Text Recognition (HTR) with limited labeled data remains a challenging problem, particularly for Arabic-script languages. Although modern sequence-based recognizers perform well in high-resource settings, their accuracy degrades sharply as training data becomes scarce. Arabic-script languages share a common writing system with substantial character overlap, motivating cross-lingual learning as a strategy to mitigate data scarcity. We conduct a controlled line-level study of cross-lingual joint training for Arabic-script HTR under low-resource regimes (number of samples K = 100, 500, 1000 labeled lines) on Arabic (KHATT), Urdu (NUST-UHWR) and Persian (PHTD). CRNN and Vision Transformer-based HTR-VT models are trained on the union of multiple related Arabic-script datasets to mitigate the data scarcity and are evaluated on individual target languages. Both architectures benefit from cross-language training under low-resource conditions. CRNN remains more effective under extremely limited target-language data, whereas the benefits of cross-language training for HTR-VT become less consistent as larger amounts of target-language data become available. On Persian (PHTD), joint training achieves a Character Error Rate (CER) of 9.99 , surpassing previously reported results despite not using the full available training data. On an additional Urdu dataset (UNHD), joint training reduces CER from 17.20 to 14.45.

The automated analysis of heterogeneous natural textures is frequently hindered by physical damage and data loss, presenting a significant challenge to computer vision. While deep learning has shown success in controlled environments, its application to complex geological materials under conditions of incomplete information remains underexplored. This study presents an integrated framework for the inpainting and classification of high-resolution core sample images. We propose an end-to-end pipeline that utilizes object detection for sample segmentation, followed by image inpainting using Generative Adversarial Networks (GANs) with Contextual Residual Aggregation (CRA) to reconstruct missing high-frequency details. Subsequently, we evaluate the performance of modern Transformer-based (Swin, ViT) and CNN architectures on the reconstructed data. Our experiments revealed a critical divergence between reconstruction quality and downstream utility: despite high structural fidelity (PSNR 28.7~dB, FID 74.01), classification accuracy plateaued at 53\%. To improve minority-class detection, we propose a confidence-based hybrid ensemble that raises MCA from 48\% to 58\%. These results highlight the limitations of current state-of-the-art generative models, which may produce visually plausible but semantically ambiguous features ("hallucinations") that confound classifiers. This work provides insights into the dependencies between image reconstruction quality and classification performance, offering a reproducible baseline for future research in non-destructive testing and material science. Given that cross-well accuracy remains in the 49--53\% range, we position the resulting system as a decision-support and screening tool for lithofacies interpretation rather than as a fully autonomous classifier. The code is available at https://github.com/GalymzhanAbdimanap/Lithology_recognition

The unstructured and irregular nature of points poses a significant challenge for accurate point cloud quality assessment (PCQA), particularly in establishing accurate perceptual feature correspondence. To tackle this, we propose the Multi-scale Implicit Structural Similarity Measurement (MS-ISSM). Unlike traditional point-to-point matching, MS-ISSM utilizes radial basis function (RBF) to represent local features continuously, transforming distortion measurement into a comparison of implicit function coefficients. This approach effectively circumvents matching errors inherent in irregular data. Additionally, we propose a ResGrouped-MLP quality assessment network, which robustly maps multi-scale feature differences to perceptual scores. The network architecture departs from traditional flat multi-layer perceptron (MLP) by adopting a grouped encoding strategy integrated with residual blocks and channel-wise attention mechanisms. This hierarchical design allows the model to preserve the distinct physical semantics of luma, chroma, and geometry while adaptively focusing on the most salient distortion features across High, Medium, and Low scales. Experimental results on multiple benchmarks demonstrate that MS-ISSM outperforms state-of-the-art metrics in both reliability and generalization. The source code is available at: https://github.com/ZhangChen2022/MS-ISSM.

Diffusion posterior sampling solves inverse problems by combining a pretrained diffusion prior with measurement-consistency guidance. However, full-band guidance can be unreliable at high noise levels, where clean estimates contain score-induced errors and high-frequency measurement directions are weakly identifiable. We argue that posterior guidance should expose measurement frequencies according to the instantaneous diffusion noise level. Based on this principle, we propose a posterior continuation framework that constructs a family of intermediate posteriors whose likelihood emphasizes currently reliable frequency bands and gradually returns to full-band consistency. We instantiate this framework with a stabilized sampler that combines a diffusion predictor, frequency-limited likelihood refinement, and a Haar-domain commitment rule that commits reliable coarse corrections while deferring weakly identifiable details. Across super-resolution, inpainting, and deblurring, our method achieves competitive-to-state-of-the-art restoration performance, including up to 5 dB PSNR improvement on motion deblurring over strong baselines in evaluations on FFHQ and ImageNet.

Generative Image Restoration (GIR) has achieved impressive perceptual realism, but how far have its practical capabilities truly advanced compared with previous methods? To answer this, we present a large-scale study grounded in a new multi-dimensional evaluation pipeline that assesses models on detail, sharpness, semantic correctness, and overall quality. Our analysis covers diverse architectures, including diffusion-based, GAN-based, PSNR-oriented, and general-purpose generation models, revealing critical performance disparities. Furthermore, our analysis uncovers a key evolution in failure modes that signifies a paradigm shift for the perception-oriented low-level vision field. The central challenge is evolving from the previous problem of detail scarcity (under-generation) to the new frontier of detail quality and semantic control (preventing over-generation). We also leverage our benchmark to train a new IQA model that better aligns with human perceptual judgments. Ultimately, this work provides a systematic study of modern generative image restoration models, offering crucial insights that redefine our understanding of their true state and chart a course for future development.

The inherent electronic and speckle noise complicates clinical interpretation of ultrasound images. Conventional denoising methods rely on explicit noise assumptions whose validity diminishes under composite noise conditions. Learning-based methods are usually pretrained in a limited image domain using a labeled dataset, which implies inevitable domain shift in complex in vivo environments. This study proposes a Pyramid Self-Contrastive Learning (PSCL) framework for test-time ultrasound image denoising without pretraining. Given multiple noisy samples from only one-shot imaging, PSCL disentangles anatomical similarity and noise randomness into separate pyramid latent spaces. The clean image is then decoded from the anatomy space while discarding the noise space. We first apply PSCL to synthetic aperture ultrasound (SAU), where an Aperture-to-Aperture loop serves as a self-supervised proxy task to ensure denoising fidelity. Simulation experiments, including noise levels from 0 to 30 dB and inclusion geometries from simple to complex, demonstrated improvements of 69.3% in SNR and 34.4% in CNR. The in vivo results showed 84.8% SNR and 25.7% CNR gains using only two aperture data of the heart in six echocardiographic views, liver, and kidney. PSCL delivers clear images across diverse imaging targets and configurations, paving the way for more reliable anatomical visualization without domain shift and pretraining costs.

Implicit Neural Representations (INRs) have recently shown impressive results, but their fundamental capacity, implicit biases, and scaling behavior remain poorly understood. We investigate the performance of diverse INRs across a suite of 2D and 3D real and synthetic signals with varying effective bandwidth, as well as both overfitting and generalization tasks including tomography, super-resolution, and denoising. By stratifying performance according to model size as well as signal type and bandwidth, our results shed light on how different INR and grid representations allocate their capacity. We find that, for many tasks involving dense signals, a simple regularized grid with interpolation trains faster and to higher or comparable quality than any INR with the same number of parameters. We also find limited settings -- namely fitting binary signals such as shape contours -- where INRs outperform grids, to guide future development and use of INRs towards the most advantageous applications.

While prior research on text-to-image generation has predominantly focused on biases in human depictions, demographic bias in generated objects remains relatively underexplored. We introduce SODA (Stereotyped Object Diagnostic Audit), a novel framework for systematically measuring these biases through automated attribute discovery and three standardized metrics: Base vs. Demographic Divergence (BDS), Cross-Demographic Disparity (CDS), and Visual Attribute Concentration (VAC). Applying SODA to 8,000 images across five state-of-the-art models and eight object categories (e.g., cars), we find that "neutral" prompts produce outputs most visually similar to middle-aged and White people, suggesting these groups are implicitly over-represented in model defaults. Furthermore, demographic cues trigger highly skewed stereotypical outputs: 26.6% of object-model-demographic combinations produce results where all 20 generated images share the exact same attribute value (e.g., rose gold laptops for women). Finally, prompt-level debiasing reduces inter-group disparity but paradoxically collapses within-group diversity, replacing one stereotype with another. SODA offers a practical pipeline for making these implicit associations measurable, serving as a step toward more responsible AI development.

Multimodal Large Language Models (MLLMs) are increasingly deployed in stateless systems, such as autonomous driving and robotics. This paper investigates a novel threat: Semantic-Aware Hijacking. We explore the feasibility of hijacking multiple stateless decisions simultaneously using a single universal perturbation. We introduce the Semantic-Aware Universal Perturbation (SAUP), which acts as a semantic router, "actively" perceiving input semantics and routing them to distinct, attacker-defined targets. To achieve this, we conduct theoretical and empirical analysis on the geometric properties in the latent space. Guided by these insights, we propose the Semantic-Oriented (SORT) optimization strategy and annotate a new dataset with fine-grained semantics to evaluate performance. Extensive experiments on three representative MLLMs demonstrate the fundamental feasibility of this attack, achieving a 66% attack success rate over five targets using a single frame against Qwen.

The widespread adoption of face recognition (FR) technologies raises serious privacy concerns, as facial data can be exploited without consent. To address this challenge, we propose Adv-TGD, a generative adversarial attack framework that synthesizes photorealistic faces capable of impersonating target identities and deceiving face recognition systems. Built upon Stable Diffusion v2.1, Adv-TGD performs per-sample LoRA fine-tuning conditioned on concise textual prompts to generate natural yet adversarially manipulated identities. Unlike conventional identity attack approaches, our method optimizes lightweight cross-attention adapters for each source-target pair within a fixed-timestep denoising process. Latent blending is constrained by a face-local heatmap mask to ensure spatially precise identity manipulation while preserving non-sensitive regions. We introduce a composite objective that integrates masked epsilon-MSE reconstruction, thresholded identity divergence in FR embedding space, directional feature alignment, and source-similarity suppression to balance adversarial attack and visual realism. Optionally, LLaVA-generated attribute prompts enhance fine-grained semantic details without reintroducing identity cues. Under the black-box evaluation protocol, Adv-TGD attains an average attack success rate (ASR) of 85.90% across IR152, IRSE50, MobileFace, and FaceNet, surpassing the semantic SOTA baseline Adv-CPG by 6.25 points, the diffusion-based makeup method DiffAIM by 3 points, and the noise-based P3-Mask by 16 points. Despite its strong attack efficacy, Adv-TGD preserves high visual fidelity (PSNR = 28.18 dB, SSIM = 0.981). Furthermore, we demonstrate the flexibility of our framework by successfully extending it to in-the-wild datasets (LADN), general object classification (ImageNet), and transformer-based diffusion models (FLUX.1).

Machine Unlearning (MUL) has emerged as a key mechanism for privacy protection and content regulation, yet current techniques often fail to guarantee the complete removal of sensitive information. While most existing works focus on verifying the execution of unlearning, they overlook the critical question of whether models remain robust against adversarial attempts to recover forgotten knowledge. In this work, we advocate for the principle of Robust Unlearning, which requires models to be both indistinguishable from retrained counterparts and resilient against diverse adversarial threats. To instantiate this principle, we propose a unified benchmark, RUB (Robust Unlearning Benchmark), that systematically evaluates the robustness of unlearning algorithms across classification, image-to-image reconstruction, and text-to-image synthesis. Within this framework, we introduce the Unlearning Mapping Attack (UMA) as a generalizable method to detect residual information, and demonstrate how existing attack strategies can be adapted into this framework as long as they conform to the generic UMA framework. Our experiments across discriminative and generative tasks reveal that state-of-the-art unlearning methods remain vulnerable under these evaluations, even when passing standard verification metrics. By positioning robustness as the central criterion and providing a benchmark for adversarial evaluation, we hope RUB paves the way toward more reliable and secure unlearning practices. The codebase and model checkpoints in RUB will be published.

Adversarial robustness of deep autoencoders (AEs) has received less attention than that of discriminative models, although their compressed latent representations induce ill-conditioned mappings that can amplify small input perturbations and destabilize reconstructions. Existing white-box attacks for AEs, which optimize norm-bounded adversarial perturbations to maximize reconstruction damage, often converge to suboptimal perturbations, thereby potentially overstating AE robustness. We show that this limitation is linked to vanishing adversarial loss gradients during backpropagation through ill-conditioned layers, associated with near-zero singular values in their intermediate weight matrices. To address this, we propose GRILL (Gradient Signal Restoration in Ill-Conditioned Layers), a framework designed to mitigate gradient degradation and improve the reliability of adversarial robustness evaluation in encoder-decoder architectures. GRILL is designed to mitigate adversarial gradient degradation during optimization, enabling attacks to better approximate high-distortion perturbations under fixed norm constraints. Through extensive experiments across multiple AE architectures, under both sample-specific and universal attacks, as well as standard and adaptive attack settings, we show that GRILL significantly increases attack effectiveness, thereby exposing vulnerabilities hidden by existing attack limitations. Beyond AEs, we provide preliminary evidence that modern multimodal encoder-decoder architectures exhibit similar vulnerabilities.

Edge-AI systems increasingly require real-time CNN inference under strict energy, performance, security, and privacy constraints. Approximate computing improves hardware efficiency by exploiting the error resilience of neural network workloads; however, most approximate CNN accelerators do not jointly consider secure, privacy-aware edge deployment. This paper presents SPARX, a Secure and Privacy-Aware Approximate CNN Acceleration framework integrated within a heterogeneous RV32IMC RISC-V System-on-Chip (SoC). SPARX combines a custom RISC-V instruction extension, an approximate logarithmic CNN acceleration unit, a lightweight differential-noise-based privacy engine, and a challenge-response authentication mechanism. To guide arithmetic selection, an approximation-aware decision framework is introduced that uses the Approximation Severity Index (ASI), Approximation Efficiency (AE), Quality of Approximation (QoA), Approximation Figure-of-Merit (AFOM), and Hardware Acceleration Efficiency (HAE). Evaluation across 11 state-of-the-art approximate MAC architectures identifies the Iterative Logarithmic Multiplier (ILM) as the most suitable design, achieving 51.7% area reduction, 81.5% power reduction, and 2.13x throughput improvement compared with an accurate radix-4 Booth MAC, while only reducing ResNet-20/CIFAR-10 accuracy by 2.82 percentage points. FPGA implementation on a Xilinx VC707 platform achieves 58.4 GOPS/W energy efficiency at 250 MHz, while 28-nm CMOS physical implementation validates ASIC feasibility

In real-world applications, a machine learning model is required to handle an open-set recognition (OSR), where unknown classes appear during the inference, in addition to a domain shift, where the data distribution differs between the training and inference phases. Domain generalization (DG) aims to handle the domain shift situation where the target domain of the inference phase is inaccessible during the model training. Open domain generalization (ODG) considers DG and OSR. Domain-augmented meta-learning (DAML) is a method targeting ODG; however, it has a complicated learning process. By contrast, although various DG methods have been proposed, they have not been evaluated in ODG situations. In this study, we comprehensively evaluate the existing DG methods in ODG and show that the two simple DG methods, CORrelation ALignment (CORAL) and maximum mean discrepancy (MMD), are competitive with DAML in several cases. In addition, we propose simple extensions of CORAL and MMD by introducing the techniques used in DAML, such as ensemble learning and Dirichlet mixup data augmentation. The experimental evaluation demonstrates that the extended CORAL and MMD can perform comparably to DAML with lower computational costs. This suggests that the simple DG methods and their simple extensions are strong baselines for ODG.

Multi-graph matching is a fundamental problem in computer vision. Our work is motivated by a challenging application in bioimaging, where dozens or even hundreds of 3D microscopy images of worms must be brought into correspondence. Existing datasets do not cover this large-scale regime, and virtually all existing methods are inapplicable because they assume a complete or dense problem setting. To support further research, our first contribution is a new large-scale dataset based on problem instances from bioimaging. Our second contribution is a comprehensive analysis of the two main multi-graph matching paradigms: direct and permutation synchronization-based formulations. We argue, in part by proof, that practical large-scale methods must explicitly address problem sparsity and incompleteness. Since standard permutation synchronization approaches fail in this setting, we further introduce a sparse permutation synchronization paradigm. Our final contribution is GREEDA, a general method for sparse and incomplete problems that can be instantiated across cost orders and paradigms. While our paper focuses on objective functions up to quadratic order, GREEDA is inherently generalizable to arbitrary orders. On larger, sparse instances, GREEDA outperforms competing methods in both objective value and runtime. For example, for moderately-sized problems based on 30 worm images GREEDA produces a high-quality solution within 2 minutes, whereas competitors require at least half an hour and yield far worse results. On smaller dense problems, GREEDA remains on par with leading methods while being an order of magnitude faster.

Human head detection, keypoint estimation, and 3D head model fitting are essential tasks with many applications. However, traditional real-world datasets often suffer from bias, privacy, and ethical concerns, and they have been recorded in laboratory environments, which makes it difficult for trained models to generalize. Here, we introduce \method -- a large-scale synthetic dataset generated with diffusion models for human head detection and 3D mesh estimation. Our dataset comprises over 1 million high-resolution images, each annotated with detailed 3D head meshes, facial landmarks, and bounding boxes. Using this dataset, we introduce a new model architecture capable of simultaneous head detection and head mesh reconstruction from a single image in a single step. Through extensive experimental evaluations, we demonstrate that models trained on our synthetic data achieve strong performance on real images. Furthermore, the versatility of our dataset makes it applicable across a broad spectrum of tasks, offering a general and comprehensive representation of human heads.

Avoiding the risk of undefined categorical labels using nearest neighbor interpolation overlooks the risk of exacerbating pixel level annotation errors in augmented training data. Additionally, the inherent low pass filtering effects of interpolation algorithms exacerbate the risk of degrading high frequency structural details within annotated regions of interest. To avoid these risks, the author modified convolutional neural networks data transformation functions by incorporating a modified geometric transformation function, removing reliance on nearest neighbor interpolation, and integrating a mean-based class filtering mechanism to handle undefined categorical labels with alternative interpolation algorithms. The author also implemented an offline data augmentation pipeline to generate interpolation specific augmented training data, enabling quantitative assessment of interpolation specific low pass filtering effects on augmented training data. Experimental evaluation on three medical image segmentation datasets and the XBAT+ datasets demonstrated performance gains across multiple quantitative metrics.

We present UniTalk, a novel dataset emphasizing challenging scenarios to enhance model generalization for the task of active speaker detection (ASD). Previously established benchmarks such as AVA predominantly comprise old movies and thus exhibit significant domain gaps with real-world video. In contrast, UniTalk covers diverse video types reflecting challenging real-world conditions, including underrepresented languages, noisy backgrounds, and crowded scenes, while being on par with AVA in scale. Extensive evaluations reveal that ASD remains unsolved under realistic conditions: state-of-the-art models near-perfect on AVA fail to reach saturation on UniTalk. Conversely, models trained on UniTalk generalize better to modern in-the-wild datasets including Talkies and ASW. UniTalk thus establishes a new benchmark for ASD, providing researchers with a valuable resource for developing and evaluating versatile and resilient models.

Salient object detection exemplifies data-bounded tasks where expensive pixel-precise annotations force separate model training for related subtasks like DIS and HR-SOD. We present a method that dramatically improves generalization through large-scale synthetic data generation and ambiguity-aware architecture. We introduce S3OD, a dataset of over 139,000 high-resolution images created through our multi-modal diffusion pipeline that extracts labels from diffusion and DINO-v3 features. The iterative generation framework prioritizes challenging categories based on model performance. We propose a streamlined multi-mask decoder that handles the inherent ambiguity in salient object detection by predicting multiple valid interpretations. Models trained only on synthetic data achieve 20-50% error reduction in cross-dataset generalization, while fine-tuned versions reach state-of-the-art performance across DIS and HR-SOD benchmarks.

E-commerce short videos represent a high-revenue segment of the online video industry characterized by a goal-driven format and dense multi-modal signals. Current models often struggle with these videos because existing benchmarks focus primarily on general-purpose tasks and neglect the reasoning of commercial intent. In this work, we first propose a multi-modal information density assessment framework to quantify the complexity of this domain. Our evaluation reveals that e-commerce content exhibits substantially higher density across visual, audio, and textual modalities compared to mainstream datasets, establishing a more challenging frontier for video understanding. To address this gap, we introduce E-commerce Video Ads Benchmark, which is the first benchmark specifically designed for e-commerce short video understanding. We curated 3,961 high-quality videos from Taobao covering a wide range of product categories and used a multi-agent system to generate 19,785 open-ended Q&A pairs, which consist of five distinct tasks. Finally, we develop E-VAds-R1, an RL-based reasoning model featuring a multi-grained reward design called MG-GRPO. This strategy provides smooth guidance for early exploration while creating a non-linear incentive for expert-level precision. Experimental results demonstrate that E-VAds-R1 achieves a 109.2% performance gain in commercial intent reasoning with only a few hundred training samples. Data is available at https://github.com/TaobaoTmall-AlgorithmProducts/E-VAds_Benchmark.

Regressing 3D rotations of objects from 2D images is a crucial yet challenging task, with broad applications in autonomous driving, virtual reality, and robotic control. Existing rotation regression models often rely on large amounts of labeled data for training or require additional information beyond 2D images, such as point clouds or CAD models. Therefore, exploring semi-supervised rotation regression using only a limited number of labeled 2D images is highly valuable. While recent work FisherMatch introduces semi-supervised learning to rotation regression, it suffers from rigid entropy-based pseudo-label filtering that fails to effectively distinguish between reliable and unreliable unlabeled samples. To address this limitation, we propose a hardness-aware curriculum learning framework that dynamically selects pseudo-labeled samples based on their difficulty, progressing from easy to complex examples. We introduce both multi-stage and adaptive curriculum strategies to replace fixed-threshold filtering with more flexible, hardness-aware mechanisms. Additionally, we present a novel structured data augmentation strategy specifically tailored for rotation estimation, which assembles composite images from augmented patches to introduce feature diversity while preserving critical geometric integrity. Comprehensive experiments on PASCAL3D+ and ObjectNet3D demonstrate that our method outperforms existing supervised and semi-supervised baselines, particularly in low-data regimes, validating the effectiveness of our curriculum learning framework and structured augmentation approach.

The offshore wind energy sector is expanding rapidly, increasing the need for independent, high-temporal-resolution monitoring of infrastructure deployment and operation at global scale. While Earth Observation based offshore wind infrastructure mapping has matured for spatial localization, existing open datasets lack temporally dense and semantically fine-grained information on construction and operational dynamics. We introduce a global Sentinel-1 synthetic aperture radar (SAR) time series data corpus that resolves deployment and operational phases of offshore wind infrastructure from 2016Q1 to 2025Q1. Building on an updated object detection workflow, we compile 15,606 time series at detected infrastructure locations, with overall 14,840,637 events as analysis-ready 1D SAR backscatter profiles, one profile per Sentinel-1 acquisition and location. To enable direct use and benchmarking, we release (i) the analysis ready 1D SAR profiles, (ii) event-level baseline semantic labels generated by a rule-based classifier, and (iii) an expert-annotated benchmark dataset of 553 time series with 328,657 event labels. The baseline classifier achieves a macro F1 score of 0.84 in event-wise evaluation and an area under the collapsed edit similarity-quality threshold curve (AUC) of 0.785, indicating temporal coherence. We demonstrate that the resulting corpus supports global-scale analyses of deployment dynamics, the identification of differences in regional deployment patterns, vessel interactions, and operational events, and provides a reference for developing and comparing time series classification methods for offshore wind infrastructure monitoring.

The Atlantic Forest in Brazil is a critical biodiversity hotspot, yet less than 12-15% of its original cover remains. Although monitoring forest restoration on a large scale is essential, traditional methods are limited by the impracticality of on-the-ground reporting on such a scale and by the saturation of remote-sensing indices such as NDVI. Furthermore, reforestation is a gradual process as opposed to the rapid spectral changes caused by deforestation. In this study, we examine 1,729 restoration sites in São Paulo, using satellite embeddings from the AlphaEarth Foundation's model to evaluate their effectiveness in characterising early restoration success. We introduce the concept of a 'Reference Trajectory Embedding', defining a metric of restoration success based on cosine similarity to reference sites of mature secondary forest. We observe distinct clusters in embedding space according to different land use and land cover (LULC) types, and we can identify sites with clear change vectors. However, the signal can be noisy, and embeddings may require further fine-tuning to capture and predict site metadata beyond LULC.

Accurate, spatially explicit characterization of tropical forest structure is essential for carbon accounting and ecosystem monitoring, yet most ML pipelines predict canopy-top height proxies (e.g., RH95/RH98) or AGBD as separate scalar targets, rather than learning the forest vertical structure as an ordered profile. The community lacks a ML-ready multimodal benchmark for predicting the entire GEDI RH profile jointly with AGBD, or for evaluating methods that enforce physically consistent ordering across RH percentiles. We address this with Biomazon, a 20 m multimodal benchmark dataset over the Amazon Basin that pairs GEDI RH and AGBD targets with multi-sensor predictors (Sentinel-1/2, ALOS-2 PALSAR-2, Copernicus DEM, Dynamic World LULC, and AlphaEarth embeddings) under standardized spatial splits and evaluation protocols. Using a shared encoder-decoder with task-specific heads as a baseline framework, we conduct a comprehensive ablation study of (i) backbone/model scale, (ii) modality contributions, and (iii) the use of auxiliary embeddings under standalone and fusion settings, and we report both single-target and joint-target results to quantify tradeoffs under a unified training protocol. Finally, we contextualize baseline performance through regionally aligned comparisons against existing gridded products, including GEDI L4D RH10-RH98 and AGBD, at matching temporal scale. Biomazon, together with the accompanying protocols and baseline results, establishes a reference benchmark for future work on structurally consistent RH-profile prediction and structure-biomass modeling in tropical forests.

Dataset condensation aims to construct compact datasets from real data via synthesis or selection. However, existing approaches are ill-suited for diffusion model training: synthetic data generation often yields low-fidelity samples unsuitable for authentic modeling, while real subset selection typically fails to preserve the distributional geometry required by diffusion likelihood objectives. To address this, we propose to reformulate real subset selection as a geometry-aware distribution alignment problem. By incorporating one-sided partial optimal transport, our method selectively aligns a compact subset with the full data distribution while allowing unmatched mass in low-density regions, ensuring the preserved geometric structure necessary for effective diffusion model training. To further ensure distributional fidelity, we complement geometric alignment with lightweight feature-statistics and semantic consistency regularization. An efficient two-stage discrete optimization strategy is proposed to achieve this alignment objective. Extensive experiments across diffusion variants, subset sizes, image resolutions, and training rounds show that our method achieves superior fidelity and distributional coverage in diffusion model training. Codes are available at https://github.com/2018cx/GADC.

Current automated pipelines for audio-visual Question Answering (QA) generally adopt a ``video-caption-QA'' paradigm. However, these methods typically segment videos into short clips and generate separate descriptions for audio and visual modalities. This decoupled processing severs inherent associations between sounds and their visual sources, while independent clip processing often causes inconsistent descriptions of the same entity across segments. Furthermore, coupling long-text comprehension and QA synthesis into a single step often restricts models to localized events, yielding questions lacking long-term temporal connections and deep cross-modal reasoning. To address these issues, we propose an automated data engine featuring two mechanisms: (1) \textbf{Entity-Anchored Video Scripting} transforms videos into structured scripts, comprising summaries, main entity lists, and segment-wise audio-visual descriptions. The entity list serves as a global prior to ensure cross-segment referential consistency and reconstruct audio-visual associations. (2) \textbf{Clue-Guided QA Generation} prompts models to first mine cross-segment, multimodal clues from the script, and subsequently generate QA pairs based on these high-value clues. Leveraging this pipeline, we construct the instruction-tuning dataset \textbf{OmniVideo-100K} and a human-verified test set, \textbf{OmniVideo-Test}. Fine-tuning VITA-1.5, Qwen2.5-Omni-7B and Qwen3-Omni-30B on OmniVideo-100K yields performance gains of up to 20.59% on OmniVideo-Test, demonstrating strong generalization (up to 12.64% improvements) across established benchmarks like Daily-Omni and JointAVBench.

Current multilingual evaluations for Vision-Language Models (VLMs) assume a one-to-one mapping between language and orthography, overlooking billions of users of multi-script languages. We introduce PuMVR (Punjabi Multimodal Visual Reasoning), a benchmark of 1,000 strictly parallel image-text instances across Punjabi's three active scripts: Gurmukhi, Shahmukhi, and Roman. Evaluating 10 state-of-the-art VLMs, we expose a substantial and systematic Script Gap. Models frequently solve visual tasks in one script while failing identical tasks in another, with accuracy deltas reaching 16%. Crucially, visual input boosts absolute performance uniformly yet does not close the orthographic gap. Furthermore, cross-script in-context transfer is highly brittle, exposing script-locked knowledge representation. Supported by McNemar tests across all script pairs, our findings demonstrate that current "multilingual" VLMs are not truly multi-script. We propose the Script Consistency Rate (SCR), which falls as low as 24.8% on our benchmark, as a mandatory metric for script-agnostic evaluation to ensure equitable AI access. Data and code are available at: https://github.com/prabhjotschugh/Not-Truly-Multilingual-PuMVR.

The development of Time-Series Forecasting (TSF) models is often constrained by the lack of comprehensive datasets, especially in Global Station Weather Forecasting (GSWF), where existing datasets are small, temporally short, and spatially sparse. To address this, we introduce WEATHER-5K, a large-scale observational weather dataset that better reflects real-world conditions, supporting improved model training and evaluation. While recent TSF methods perform well on benchmarks, they lag behind operational Numerical Weather Prediction systems in capturing complex weather dynamics and extreme events. We propose PhysicsFormer, a physics-informed forecasting model combining a dynamic core with a Transformer residual to predict future weather states. Physical consistency is enforced via pressure-wind alignment and energy-aware smoothness losses, ensuring plausible dynamics while capturing complex temporal patterns. We benchmark PhysicsFormer and other TSF models against operational systems across several weather variables, extreme event prediction, and model complexity, providing a comprehensive assessment of the gap between academic TSF models and operational forecasting. The dataset and benchmark implementation are available at: https://github.com/taohan10200/WEATHER-5K.

In this paper, we delve deeper into the Kullback-Leibler (KL) Divergence loss and mathematically prove that it is equivalent to the Decoupled Kullback-Leibler (DKL) Divergence loss that consists of (1) a weighted Mean Square Error (wMSE) loss and (2) a Cross-Entropy loss incorporating soft labels. Thanks to the decoupled structure of DKL loss, we have identified two areas for improvement. Firstly, we address the limitation of KL loss in scenarios like knowledge distillation by breaking its asymmetric optimization property along with a smoother weight function. This modification effectively alleviates convergence challenges in optimization, particularly for classes with high predicted scores in soft labels. Secondly, we introduce class-wise global information into KL/DKL to reduce bias arising from individual samples. With these two enhancements, we derive the Generalized Kullback-Leibler (GKL) Divergence loss and evaluate its effectiveness by conducting experiments on CIFAR-10/100, ImageNet, and vision-language datasets, focusing on adversarial training, and knowledge distillation tasks. Specifically, we achieve new state-of-the-art adversarial robustness on the public leaderboard -- RobustBench and competitive knowledge distillation performance across CIFAR/ImageNet models and CLIP models, demonstrating the substantial practical merits. Our code is available at https://github.com/jiequancui/DKL.

Generalist embodied agents require more than object recognition: they must reason about spatial relations, actions, procedures, human intentions, environmental constraints, and commonsense consequences from situated visual observations. Yet existing visual and embodied question answering benchmarks often provide limited control over the reasoning dependencies being tested, making it difficult to distinguish grounded embodied reasoning from shortcut-driven visual or linguistic pattern matching. We present ERQA-Plus, a diagnostic benchmark for reasoning in embodied AI. ERQA-Plus contains 1,766 question-answer instances grounded in 711 robot-centric images and organized according to a structured taxonomy spanning perceptual, action-centric, social-interaction, navigation-environmental, and contextual commonsense reasoning. The dataset is constructed using a multi-stage generation and validation pipeline that combines taxonomy-guided question generation, automatic quality judging, iterative revision, and human assessment to improve visual grounding, answer validity, and reasoning quality. We benchmark representative general-purpose vision-language models and embodied models, including LLaVA-NeXT-8B, Prismatic-7B, MiniCPM-V-4.5-8B, Qwen3-VL, RoboRefer-8B, and RoboBrain2.5-8B. Although the strongest model, Qwen3-VL-32B, achieves 83.4% overall accuracy and 61.4 SBERT score, category-level results reveal persistent weaknesses in spatial reasoning, procedural reasoning, event prediction, and intention inference. ERQA-Plus therefore provides a fine-grained evaluation framework for measuring not only whether embodied agents answer correctly, but also which forms of embodied reasoning they can and cannot perform reliably. The dataset is available https://huggingface.co/datasets/huggingdas/erqa-plus and the project page at https://github.com/LUNAProject22/erqa-plus.

Even during fixation the human eye is constantly in low amplitude motion, jittering over small angles in random directions at up to 100Hz. This motion results in all features of the image on the retina constantly traversing a number of cones, yet objects which are stable in the world are perceived to be stable, and any object which is moving in the world is perceived to be moving. A series of experiments carried out over a dozen years revealed the psychophysics of visual stabilization to be more nuanced than might be assumed, say, from the mechanics of stabilization of camera images, or what might be assumed to be the simplest solution from an evolutionary perspective. The psychophysics revealed by the experiments strongly implies a specific set of operations on retinal signals resulting in the observed stabilization behavior. The presentation is in two levels. First is a functional description of the action of the mechanism that is very likely responsible for the experimentally observed behavior. Second is a more speculative proposal of circuit-level neural elements that might implement the functional behavior.

Geometric problem solving, as a typical multimodal reasoning problem, has attracted much attention and made great progress recently, however most of works focus on plane geometry while usually fail in solid geometry due to 3D spatial diagrams and complex reasoning. To bridge this gap, we introduce Hilbert-Geo, the first unified formal language framework for solid geometry, including an extensive predicate library and a dedicated theorem bank. Based on this framework, we propose a Parse2Reason method containing two steps of first parsing then reasoning. In the parsing step, we utilize conditional description language (CDL), a formalized language composed of predicates specifically designed to construct geometric conditions, to represent both problem description (natural text) and solid diagrams (visual image). In the reasoning step, we leverage those formal CDL and the theorem bank to perform relational inference and algebraic computation, generating strictly correct, verifiable, and human-readable reasoning processes. Notably, our proposed Hilbert-Geo is also applicable to plane geometry. To advance geometric reasoning, we curate two expert-annotated dataset SolidFGeo2k and PlaneFGeo3k, which are furnished with geometric formal language annotations, solutions and answers. Extensive experiments show that our proposed method achieves the state-of-the-art (SOTA) performance 77.3% in SolidFGeo2k and 84.1% in MathVerse-Solid (one small subset in MathVerse dedicated to solid geometry), substantially outperforming leading MLLMs, such as Gemini-2.5-pro (54.2% on SolidFGeo2k) and GPT-5 (62.9% on MathVerse-Solid). In addition, our method achieves the SOTA accuracy 80.2% in PlaneFGeo3k, demonstrating the generality of the Hilbert-Geo in geometric reasoning. Our code and datasets are released at https://github.com/PremiLab-Math/Hilbert-Geo.

Zero-shot text-to-speech (TTS) has made significant progress in replicating unseen voices, yet balancing generation quality and inference efficiency remains challenging. Autoregressive models suffer from high latency, while diffusion-based approaches are constrained by training-time configurations. Moreover, most flow-based methods operate in continuous space, which introduces optimization challenges because continuous token spaces are inherently more complex than discrete ones. To address these limitations, we propose DiFlow-TTS, a novel zero-shot TTS framework based on discrete flow matching. The model consists of a deterministic Phoneme-Content Mapper for linguistic modeling and a Factorized Discrete Flow Denoiser that simultaneously generates prosody and acoustic token streams. Experimental results demonstrate the effectiveness of our approach across multiple evaluation metrics.

Alzheimer's disease (AD) confirmation often relies on positron emission tomography (PET) or cerebrospinal fluid (CSF) analysis, which are costly and invasive. Consequently, structural MRI biomarkers such as cortical thickness (CT) are widely used for non-invasive AD screening. Multiscale structural mapping (MSSM) was recently proposed to integrate gray-white matter contrasts (GWCs) with CT from a single T1-weighted MRI (T1w) scan. Building on this framework, we propose MSSM+, together with surface supervertex mapping (SSVM) and a Supervertex Vision Transformer (SV-ViT). 3D T1w images from individuals with AD and cognitively normal (CN) controls were analyzed. MSSM+ extends MSSM by incorporating sulcal depth and cortical curvature at the vertex level. SSVM partitions the cortical surface into supervertices (surface patches) that effectively represent inter- and intra-regional spatial relationships. SV-ViT is a Vision Transformer architecture operating on these supervertices, enabling anatomically informed learning from surface mesh representations. Compared with MSSM, MSSM+ identified more spatially extensive and statistically significant group differences between AD and CN. In AD vs. CN classification, MSSM+ achieved a 3%p higher area under the precision-recall curve than MSSM. Vendor-specific analyses further demonstrated reduced signal variability and consistently improved classification performance across MR manufacturers relative to CT, GWCs, and MSSM. These findings suggest that MSSM+ combined with SV-ViT is a promising MRI-based imaging marker for AD detection prior to CSF/PET confirmation.

Accurate histopathologic interpretation is key for clinical decision-making; however, current deep learning models for digital pathology are often overconfident and poorly calibrated in out-of-distribution (OOD) settings, which limit trust and clinical adoption. Safety-critical medical imaging workflows benefit from intrinsic uncertainty-aware properties that can accurately reject OOD input. We implement the Spectral-normalized Neural Gaussian Process (SNGP), a set of lightweight modifications that apply spectral normalization and replace the final dense layer with a Gaussian process layer to improve single-model uncertainty estimation and OOD detection. We evaluate SNGP vs. deterministic and MonteCarlo dropout on six datasets across three biomedical classification tasks: white blood cells, amyloid plaques, and colorectal histopathology. SNGP has comparable in-distribution performance while significantly improving uncertainty estimation and OOD detection. Thus, SNGP or related models offer a useful framework for uncertainty-aware classification in digital pathology, supporting safe deployment and building trust with pathologists.

Episodic memory retrieval enables wearable cameras to recall objects or events previously observed in video. However, existing formulations assume an "offline" setting with full video access at query time, limiting their applicability in real-world scenarios with power and storage-constrained wearable devices. Towards more application-ready episodic memory systems, we introduce Online Visual Query 2D (OVQ2D), a task where models process video streams online, observing each frame only once, and retrieve object localizations using a compact memory instead of full video history. We address OVQ2D with ESOM (Egocentric Streaming Object Memory), a novel framework integrating an object discovery module, an object tracking module, and a memory module that find, track, and store spatio-temporal object information for efficient querying. Experiments on Ego4D demonstrate ESOM's superiority over other online approaches, though OVQ2D remains challenging, with top performance at only ~4% success. ESOM's accuracy increases markedly with perfect object tracking (31.91%), discovery (40.55%), or both (81.92%), underscoring the need of applied research on these components.

In positron emission tomography (PET), it is indispensable to perform attenuation correction in order to obtain the quantitatively accurate activity map (tracer distribution) in the body. Generally, this is carried out based on the estimated attenuation map obtained from computed tomography or magnetic resonance imaging. However, except for errors in the attenuation correction factors obtained, the additional scan not only brings in new radiation doses and/or increases the scanning time but also leads to severe misalignment induced by various motions during and between the two sequential scans. To address these issues, based on maximum likelihood estimation, we propose a new mathematical model for simultaneously reconstructing the activity and attenuation sinogram from the time-of-flight (TOF)-PET emission data only. Particularly, we make full use of the exclusively exponential form for the attenuation correction factors, and consider the constraint of a total amount of the activity in some mask region in the proposed model. Furthermore, we prove its well-posedness, including the existence, uniqueness and stability of the solution. We propose an alternating update algorithm to solve the model, and also analyze its convergence. Finally, numerical experiments with various TOF-PET emission data demonstrate that the proposed method is of numerical convergence and robust to noise, and outperforms some state-of-the-art methods in terms of accuracy and efficiency, and has the capability of autonomous attenuation correction.

Limited-Angle Computed Tomography (LACT) is a challenging inverse problem where missing angular projections lead to incomplete sinograms and severe artifacts in the reconstructed images. While recent learning-based methods have demonstrated effectiveness, most of them assume ideal, noise-free measurements and fail to address the impact of measurement noise. To overcome this limitation, we treat LACT as a sinogram inpainting task and propose a diffusion-based framework that completes missing angular views using a Mean-Reverting Stochastic Differential Equation (MR-SDE) formulation. To improve robustness under realistic noise, we propose RNSD$^+$, a novel noise-aware rectification mechanism that explicitly models inference-time uncertainty, enabling reliable and robust reconstruction. Extensive experiments demonstrate that our method consistently surpasses baseline models in data consistency and perceptual quality, and generalizes well across varying noise intensity and acquisition scenarios.

Real-world application models are commonly deployed in dynamic environments, where the target domain distribution undergoes temporal changes. Continual Test-Time Adaptation (CTTA) has recently emerged as a promising technique to gradually adapt a source-trained model to continually changing target domains. Despite recent advancements in addressing CTTA, two critical issues remain: 1) Fixed thresholds for pseudo-labeling in existing methodologies lead to low-quality pseudo-labels, as model confidence varies across categories and domains; 2) Stochastic parameter restoration methods for mitigating catastrophic forgetting fail to preserve critical information effectively, due to their intrinsic randomness. To tackle these challenges for detection models in CTTA scenarios, we present AMROD, featuring three core components. Firstly, the object-level contrastive learning module extracts object-level features for contrastive learning to refine the feature representation in the target domain. Secondly, the adaptive monitoring module dynamically skips unnecessary adaptation and updates the category-specific threshold based on predicted confidence scores to enable efficiency and improve the quality of pseudo-labels. Lastly, the adaptive randomized restoration mechanism selectively reset inactive parameters with higher possibilities, ensuring the retention of essential knowledge. We demonstrate the effectiveness of AMROD on four CTTA object detection tasks, where AMROD outperforms existing methods, especially achieving a 3.2 mAP improvement and a 20% increase in efficiency on the Cityscapes-to-Cityscapes-C CTTA task. The code of this work is available at https://github.com/ShileiCao/AMROD.