Dingkang Yang

Yue Jiang, Xue Jiang, Lihua Zhang et al.

Multimodal large language models (MLLMs) demonstrate remarkable visual understanding, yet their reliability in interactive settings is severely undermined by hallucination snowballing: a phenomenon where initial errors amplify across conversational turns, leading to a collapse in coherence. This failure reveals a fundamental vulnerability where models progressively neglect visual grounding in favor of over-relying on polluted textual history. Existing benchmarks are predominantly confined to single-turn VQA, which fail to capture the complex dynamics of error propagation in long-horizon interactions. To address this, we introduce MM-Snowball, the first benchmark for fine-grained diagnosis of hallucination snowballing within dialogues. Extensive evaluation shows that our benchmark poses a significant challenge even to advanced MLLMs and reveals the inefficacy of existing mitigation methods designed for single-turn VQA. To counteract this degradation, we propose Conflict-Aware Visual Rectification (CAVR). This training-free method mitigates snowballing through a synergistic dual-mechanism that refreshes visual grounding at the representation level and rectifies output distributions at the logit level, effectively re-anchoring the model to visual facts. Experiments demonstrate that CAVR achieves state-of-the-art performance, offering a promising path toward more reliable interactive AI. Data and code are available at: https://frenkie-chiang.github.io/MM-Snowball

Dingkang Yang, Shuai Huang, Zhi Xu et al.

Driver distraction has become a significant cause of severe traffic accidents over the past decade. Despite the growing development of vision-driven driver monitoring systems, the lack of comprehensive perception datasets restricts road safety and traffic security. In this paper, we present an AssIstive Driving pErception dataset (AIDE) that considers context information both inside and outside the vehicle in naturalistic scenarios. AIDE facilitates holistic driver monitoring through three distinctive characteristics, including multi-view settings of driver and scene, multi-modal annotations of face, body, posture, and gesture, and four pragmatic task designs for driving understanding. To thoroughly explore AIDE, we provide experimental benchmarks on three kinds of baseline frameworks via extensive methods. Moreover, two fusion strategies are introduced to give new insights into learning effective multi-stream/modal representations. We also systematically investigate the importance and rationality of the key components in AIDE and benchmarks. The project link is https://github.com/ydk122024/AIDE.

Jiafeng Wang, Zhaoyu Chen, Kaixun Jiang et al.

Deep Neural Networks (DNNs) are vulnerable to adversarial examples, which are crafted by adding human-imperceptible perturbations to the benign inputs. Simultaneously, adversarial examples exhibit transferability across models, enabling practical black-box attacks. However, existing methods are still incapable of achieving the desired transfer attack performance. In this work, focusing on gradient optimization and consistency, we analyse the gradient elimination phenomenon as well as the local momentum optimum dilemma. To tackle these challenges, we introduce Global Momentum Initialization (GI), providing global momentum knowledge to mitigate gradient elimination. Specifically, we perform gradient pre-convergence before the attack and a global search during this stage. GI seamlessly integrates with existing transfer methods, significantly improving the success rate of transfer attacks by an average of 6.4% under various advanced defense mechanisms compared to the state-of-the-art method. Ultimately, GI demonstrates strong transferability in both image and video attack domains. Particularly, when attacking advanced defense methods in the image domain, it achieves an average attack success rate of 95.4%. The code is available at $\href{https://github.com/Omenzychen/Global-Momentum-Initialization}{https://github.com/Omenzychen/Global-Momentum-Initialization}$.

Jiyuan Fu, Kaixun Jiang, Jingkai Jia et al.

While Vision-Language-Action (VLA) models have emerged as powerful generalist policies, their severe vulnerability to adversarial patches significantly hinders their deployment in safety-critical domains. Moreover, existing patch attacks primarily focus on white-box settings, heavily overfitting to the specific action output space of the target model, which results in poor cross-architecture transferability. To overcome this limitation, we propose VLA-Hijack, a unified adversarial framework that breaks the transferability bottleneck by exploiting a fundamental vulnerability identified in this work: before planning any motion, a VLA model must first use visual information to locate its own robotic arm within the environment. Targeting this shared visual self-localization process, our approach concurrently optimizes Attention-Guided Proprioceptive Suppression to inhibit the real robotic arm's features, and Multimodal Proprioceptive Injection to establish the patch as a surrogate "phantom embodiment". By alternating between semantic concept anchoring and visual prototype projection, VLA-Hijack effectively severs the semantic relationship between the agent's true embodiment and its control policy. Extensive experiments across diverse architectures (OpenVLA, UniVLA, and CronusVLA) demonstrate that VLA-Hijack achieves superior optimization efficiency in white-box settings and sets a new SOTA for cross-architecture and cross-domain black-box transferability.

Shunli Wang, Shuaibing Wang, Bo Jiao et al.

Reliable and stable 6D pose estimation of uncooperative space objects plays an essential role in on-orbit servicing and debris removal missions. Considering that the pose estimator is sensitive to background interference, this paper proposes a counterfactual analysis framework named CASpaceNet to complete robust 6D pose estimation of the spaceborne targets under complicated background. Specifically, conventional methods are adopted to extract the features of the whole image in the factual case. In the counterfactual case, a non-existent image without the target but only the background is imagined. Side effect caused by background interference is reduced by counterfactual analysis, which leads to unbiased prediction in final results. In addition, we also carry out lowbit-width quantization for CA-SpaceNet and deploy part of the framework to a Processing-In-Memory (PIM) accelerator on FPGA. Qualitative and quantitative results demonstrate the effectiveness and efficiency of our proposed method. To our best knowledge, this paper applies causal inference and network quantization to the 6D pose estimation of space-borne targets for the first time. The code is available at https://github.com/Shunli-Wang/CA-SpaceNet.

Minghao Han, Linhao Qu, Dingkang Yang et al.

Multiple instance learning (MIL) has become a standard paradigm for the weakly supervised classification of whole slide images (WSIs). However, this paradigm relies on using a large number of labeled WSIs for training. The lack of training data and the presence of rare diseases pose significant challenges for these methods. Prompt tuning combined with pre-trained Vision-Language models (VLMs) is an effective solution to the Few-shot Weakly Supervised WSI Classification (FSWC) task. Nevertheless, applying prompt tuning methods designed for natural images to WSIs presents three significant challenges: 1) These methods fail to fully leverage the prior knowledge from the VLM's text modality; 2) They overlook the essential multi-scale and contextual information in WSIs, leading to suboptimal results; and 3) They lack exploration of instance aggregation methods. To address these problems, we propose a Multi-Scale and Context-focused Prompt Tuning (MSCPT) method for FSWC task. Specifically, MSCPT employs the frozen large language model to generate pathological visual language prior knowledge at multiple scales, guiding hierarchical prompt tuning. Additionally, we design a graph prompt tuning module to learn essential contextual information within WSI, and finally, a non-parametric cross-guided instance aggregation module has been introduced to derive the WSI-level features. Extensive experiments, visualizations, and interpretability analyses were conducted on five datasets and three downstream tasks using three VLMs, demonstrating the strong performance of our MSCPT. All codes have been made publicly accessible at https://github.com/Hanminghao/MSCPT.

Yang Liu, Dingkang Yang, Yan Wang et al.

Video Anomaly Detection (VAD) serves as a pivotal technology in the intelligent surveillance systems, enabling the temporal or spatial identification of anomalous events within videos. While existing reviews predominantly concentrate on conventional unsupervised methods, they often overlook the emergence of weakly-supervised and fully-unsupervised approaches. To address this gap, this survey extends the conventional scope of VAD beyond unsupervised methods, encompassing a broader spectrum termed Generalized Video Anomaly Event Detection (GVAED). By skillfully incorporating recent advancements rooted in diverse assumptions and learning frameworks, this survey introduces an intuitive taxonomy that seamlessly navigates through unsupervised, weakly-supervised, supervised and fully-unsupervised VAD methodologies, elucidating the distinctions and interconnections within these research trajectories. In addition, this survey facilitates prospective researchers by assembling a compilation of research resources, including public datasets, available codebases, programming tools, and pertinent literature. Furthermore, this survey quantitatively assesses model performance, delves into research challenges and directions, and outlines potential avenues for future exploration.

Kun Yang, Dingkang Yang, Jingyu Zhang et al.

Multi-agent collaborative perception as a potential application for vehicle-to-everything communication could significantly improve the perception performance of autonomous vehicles over single-agent perception. However, several challenges remain in achieving pragmatic information sharing in this emerging research. In this paper, we propose SCOPE, a novel collaborative perception framework that aggregates the spatio-temporal awareness characteristics across on-road agents in an end-to-end manner. Specifically, SCOPE has three distinct strengths: i) it considers effective semantic cues of the temporal context to enhance current representations of the target agent; ii) it aggregates perceptually critical spatial information from heterogeneous agents and overcomes localization errors via multi-scale feature interactions; iii) it integrates multi-source representations of the target agent based on their complementary contributions by an adaptive fusion paradigm. To thoroughly evaluate SCOPE, we consider both real-world and simulated scenarios of collaborative 3D object detection tasks on three datasets. Extensive experiments demonstrate the superiority of our approach and the necessity of the proposed components.

Yang Liu, Jing Liu, Mengyang Zhao et al.

Video anomaly detection is a challenging task in the computer vision community. Most single task-based methods do not consider the independence of unique spatial and temporal patterns, while two-stream structures lack the exploration of the correlations. In this paper, we propose spatial-temporal memories augmented two-stream auto-encoder framework, which learns the appearance normality and motion normality independently and explores the correlations via adversarial learning. Specifically, we first design two proxy tasks to train the two-stream structure to extract appearance and motion features in isolation. Then, the prototypical features are recorded in the corresponding spatial and temporal memory pools. Finally, the encoding-decoding network performs adversarial learning with the discriminator to explore the correlations between spatial and temporal patterns. Experimental results show that our framework outperforms the state-of-the-art methods, achieving AUCs of 98.1% and 89.8% on UCSD Ped2 and CUHK Avenue datasets.

Kun Yang, Jing Liu, Dingkang Yang et al.

With the rapid development of intelligent transportation system applications, a tremendous amount of multi-view video data has emerged to enhance vehicle perception. However, performing video analytics efficiently by exploiting the spatial-temporal redundancy from video data remains challenging. Accordingly, we propose a novel traffic-related framework named CEVAS to achieve efficient object detection using multi-view video data. Briefly, a fine-grained input filtering policy is introduced to produce a reasonable region of interest from the captured images. Also, we design a sharing object manager to manage the information of objects with spatial redundancy and share their results with other vehicles. We further derive a content-aware model selection policy to select detection methods adaptively. Experimental results show that our framework significantly reduces response latency while achieving the same detection accuracy as the state-of-the-art methods.

Zhongying Deng, Cheng Tang, Ziyan Huang et al. · pku

Foundation models have demonstrated remarkable success across diverse domains and tasks, primarily due to the thrive of large-scale, diverse, and high-quality datasets. However, in the field of medical imaging, the curation and assembling of such medical datasets are highly challenging due to the reliance on clinical expertise and strict ethical and privacy constraints, resulting in a scarcity of large-scale unified medical datasets and hindering the development of powerful medical foundation models. In this work, we present the largest survey to date of medical image datasets, covering over 1,000 open-access datasets with a systematic catalog of their modalities, tasks, anatomies, annotations, limitations, and potential for integration. Our analysis exposes a landscape that is modest in scale, fragmented across narrowly scoped tasks, and unevenly distributed across organs and modalities, which in turn limits the utility of existing medical image datasets for developing versatile and robust medical foundation models. To turn fragmentation into scale, we propose a metadata-driven fusion paradigm (MDFP) that integrates public datasets with shared modalities or tasks, thereby transforming multiple small data silos into larger, more coherent resources. Building on MDFP, we release an interactive discovery portal that enables end-to-end, automated medical image dataset integration, and compile all surveyed datasets into a unified, structured table that clearly summarizes their key characteristics and provides reference links, offering the community an accessible and comprehensive repository. By charting the current terrain and offering a principled path to dataset consolidation, our survey provides a practical roadmap for scaling medical imaging corpora, supporting faster data discovery, more principled dataset creation, and more capable medical foundation models.

Kaixun Jiang, Zhaoyu Chen, Hao Huang et al.

Although Deep Neural Networks (DNNs) have demonstrated excellent performance, they are vulnerable to adversarial patches that introduce perceptible and localized perturbations to the input. Generating adversarial patches on images has received much attention, while adversarial patches on videos have not been well investigated. Further, decision-based attacks, where attackers only access the predicted hard labels by querying threat models, have not been well explored on video models either, even if they are practical in real-world video recognition scenes. The absence of such studies leads to a huge gap in the robustness assessment for video models. To bridge this gap, this work first explores decision-based patch attacks on video models. We analyze that the huge parameter space brought by videos and the minimal information returned by decision-based models both greatly increase the attack difficulty and query burden. To achieve a query-efficient attack, we propose a spatial-temporal differential evolution (STDE) framework. First, STDE introduces target videos as patch textures and only adds patches on keyframes that are adaptively selected by temporal difference. Second, STDE takes minimizing the patch area as the optimization objective and adopts spatialtemporal mutation and crossover to search for the global optimum without falling into the local optimum. Experiments show STDE has demonstrated state-of-the-art performance in terms of threat, efficiency and imperceptibility. Hence, STDE has the potential to be a powerful tool for evaluating the robustness of video recognition models.

Minghao Han, Yichen Liu, Yizhou Liu et al.

In computational pathology, understanding and generation have evolved along disparate paths: advanced understanding models already exhibit diagnostic-level competence, whereas generative models largely simulate pixels. Progress remains hindered by three coupled factors: the scarcity of large, high-quality image-text corpora; the lack of precise, fine-grained semantic control, which forces reliance on non-semantic cues; and terminological heterogeneity, where diverse phrasings for the same diagnostic concept impede reliable text conditioning. We introduce UniPath, a semantics-driven pathology image generation framework that leverages mature diagnostic understanding to enable controllable generation. UniPath implements Multi-Stream Control: a Raw-Text stream; a High-Level Semantics stream that uses learnable queries to a frozen pathology MLLM to distill paraphrase-robust Diagnostic Semantic Tokens and to expand prompts into diagnosis-aware attribute bundles; and a Prototype stream that affords component-level morphological control via a prototype bank. On the data front, we curate a 2.65M image-text corpus and a finely annotated, high-quality 68K subset to alleviate data scarcity. For a comprehensive assessment, we establish a four-tier evaluation hierarchy tailored to pathology. Extensive experiments demonstrate UniPath's SOTA performance, including a Patho-FID of 80.9 (51% better than the second-best) and fine-grained semantic control achieving 98.7% of the real-image. The dataset and code can be obtained from https://github.com/Hanminghao/UniPath.

Yuzheng Wang, Zhaoyu Chen, Jie Zhang et al.

Data-Free Knowledge Distillation (DFKD) is a novel task that aims to train high-performance student models using only the pre-trained teacher network without original training data. Most of the existing DFKD methods rely heavily on additional generation modules to synthesize the substitution data resulting in high computational costs and ignoring the massive amounts of easily accessible, low-cost, unlabeled open-world data. Meanwhile, existing methods ignore the domain shift issue between the substitution data and the original data, resulting in knowledge from teachers not always trustworthy and structured knowledge from data becoming a crucial supplement. To tackle the issue, we propose a novel Open-world Data Sampling Distillation (ODSD) method for the DFKD task without the redundant generation process. First, we try to sample open-world data close to the original data's distribution by an adaptive sampling module and introduce a low-noise representation to alleviate the domain shift issue. Then, we build structured relationships of multiple data examples to exploit data knowledge through the student model itself and the teacher's structured representation. Extensive experiments on CIFAR-10, CIFAR-100, NYUv2, and ImageNet show that our ODSD method achieves state-of-the-art performance with lower FLOPs and parameters. Especially, we improve 1.50\%-9.59\% accuracy on the ImageNet dataset and avoid training the separate generator for each class.

Yuzheng Wang, Zhaoyu Chen, Dingkang Yang et al.

Adversarial Robustness Distillation (ARD) is a novel method to boost the robustness of small models. Unlike general adversarial training, its robust knowledge transfer can be less easily restricted by the model capacity. However, the teacher model that provides the robustness of knowledge does not always make correct predictions, interfering with the student's robust performances. Besides, in the previous ARD methods, the robustness comes entirely from one-to-one imitation, ignoring the relationship between examples. To this end, we propose a novel structured ARD method called Contrastive Relationship DeNoise Distillation (CRDND). We design an adaptive compensation module to model the instability of the teacher. Moreover, we utilize the contrastive relationship to explore implicit robustness knowledge among multiple examples. Experimental results on multiple attack benchmarks show CRDND can transfer robust knowledge efficiently and achieves state-of-the-art performances.

Dingkang Yang, Zhaoyu Chen, Yuzheng Wang et al.

Context-Aware Emotion Recognition (CAER) is a crucial and challenging task that aims to perceive the emotional states of the target person with contextual information. Recent approaches invariably focus on designing sophisticated architectures or mechanisms to extract seemingly meaningful representations from subjects and contexts. However, a long-overlooked issue is that a context bias in existing datasets leads to a significantly unbalanced distribution of emotional states among different context scenarios. Concretely, the harmful bias is a confounder that misleads existing models to learn spurious correlations based on conventional likelihood estimation, significantly limiting the models' performance. To tackle the issue, this paper provides a causality-based perspective to disentangle the models from the impact of such bias, and formulate the causalities among variables in the CAER task via a tailored causal graph. Then, we propose a Contextual Causal Intervention Module (CCIM) based on the backdoor adjustment to de-confound the confounder and exploit the true causal effect for model training. CCIM is plug-in and model-agnostic, which improves diverse state-of-the-art approaches by considerable margins. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our CCIM and the significance of causal insight.

Siao Liu, Zhaoyu Chen, Yang Liu et al.

Learning a policy with great generalization to unseen environments remains challenging but critical in visual reinforcement learning. Despite the success of augmentation combination in the supervised learning generalization, naively applying it to visual RL algorithms may damage the training efficiency, suffering from serve performance degradation. In this paper, we first conduct qualitative analysis and illuminate the main causes: (i) high-variance gradient magnitudes and (ii) gradient conflicts existed in various augmentation methods. To alleviate these issues, we propose a general policy gradient optimization framework, named Conflict-aware Gradient Agreement Augmentation (CG2A), and better integrate augmentation combination into visual RL algorithms to address the generalization bias. In particular, CG2A develops a Gradient Agreement Solver to adaptively balance the varying gradient magnitudes, and introduces a Soft Gradient Surgery strategy to alleviate the gradient conflicts. Extensive experiments demonstrate that CG2A significantly improves the generalization performance and sample efficiency of visual RL algorithms.

Shunli Wang, Dingkang Yang, Peng Zhai et al.

Human action recognition and analysis have great demand and important application significance in video surveillance, video retrieval, and human-computer interaction. The task of human action quality evaluation requires the intelligent system to automatically and objectively evaluate the action completed by the human. The action quality assessment model can reduce the human and material resources spent in action evaluation and reduce subjectivity. In this paper, we provide a comprehensive survey of existing papers on video-based action quality assessment. Different from human action recognition, the application scenario of action quality assessment is relatively narrow. Most of the existing work focuses on sports and medical care. We first introduce the definition and challenges of human action quality assessment. Then we present the existing datasets and evaluation metrics. In addition, we summarized the methods of sports and medical care according to the model categories and publishing institutions according to the characteristics of the two fields. At the end, combined with recent work, the promising development direction in action quality assessment is discussed.

Linhao Qu, Dingkang Yang, Dan Huang et al.

Current multi-instance learning algorithms for pathology image analysis often require a substantial number of Whole Slide Images for effective training but exhibit suboptimal performance in scenarios with limited learning data. In clinical settings, restricted access to pathology slides is inevitable due to patient privacy concerns and the prevalence of rare or emerging diseases. The emergence of the Few-shot Weakly Supervised WSI Classification accommodates the significant challenge of the limited slide data and sparse slide-level labels for diagnosis. Prompt learning based on the pre-trained models (\eg, CLIP) appears to be a promising scheme for this setting; however, current research in this area is limited, and existing algorithms often focus solely on patch-level prompts or confine themselves to language prompts. This paper proposes a multi-instance prompt learning framework enhanced with pathology knowledge, \ie, integrating visual and textual prior knowledge into prompts at both patch and slide levels. The training process employs a combination of static and learnable prompts, effectively guiding the activation of pre-trained models and further facilitating the diagnosis of key pathology patterns. Lightweight Messenger (self-attention) and Summary (attention-pooling) layers are introduced to model relationships between patches and slides within the same patient data. Additionally, alignment-wise contrastive losses ensure the feature-level alignment between visual and textual learnable prompts for both patches and slides. Our method demonstrates superior performance in three challenging clinical tasks, significantly outperforming comparative few-shot methods.

Dingkang Yang, Mingcheng Li, Linhao Qu et al.

Understanding human intentions (e.g., emotions) from videos has received considerable attention recently. Video streams generally constitute a blend of temporal data stemming from distinct modalities, including natural language, facial expressions, and auditory clues. Despite the impressive advancements of previous works via attention-based paradigms, the inherent temporal asynchrony and modality heterogeneity challenges remain in multimodal sequence fusion, causing adverse performance bottlenecks. To tackle these issues, we propose a Multimodal fusion approach for learning modality-Exclusive and modality-Agnostic representations (MEA) to refine multimodal features and leverage the complementarity across distinct modalities. On the one hand, MEA introduces a predictive self-attention module to capture reliable context dynamics within modalities and reinforce unique features over the modality-exclusive spaces. On the other hand, a hierarchical cross-modal attention module is designed to explore valuable element correlations among modalities over the modality-agnostic space. Meanwhile, a double-discriminator strategy is presented to ensure the production of distinct representations in an adversarial manner. Eventually, we propose a decoupled graph fusion mechanism to enhance knowledge exchange across heterogeneous modalities and learn robust multimodal representations for downstream tasks. Numerous experiments are implemented on three multimodal datasets with asynchronous sequences. Systematic analyses show the necessity of our approach.

Haopeng Kuang, Dingkang Yang, Shunli Wang et al.

Accurate visualization of liver tumors and their surrounding blood vessels is essential for noninvasive diagnosis and prognosis prediction of tumors. In medical image segmentation, there is still a lack of in-depth research on the simultaneous segmentation of liver tumors and peritumoral blood vessels. To this end, we collect the first liver tumor, and vessel segmentation benchmark datasets containing 52 portal vein phase computed tomography images with liver, liver tumor, and vessel annotations. In this case, we propose a 3D U-shaped Cross-Attention Network (UCA-Net) that utilizes a tailored cross-attention mechanism instead of the traditional skip connection to effectively model the encoder and decoder feature. Specifically, the UCA-Net uses a channel-wise cross-attention module to reduce the semantic gap between encoder and decoder and a slice-wise cross-attention module to enhance the contextual semantic learning ability among distinct slices. Experimental results show that the proposed UCA-Net can accurately segment 3D medical images and achieve state-of-the-art performance on the liver tumor and intrahepatic vessel segmentation task.

Yuxuan Lei, Dingkang Yang, Zhaoyu Chen et al.

Context-aware emotion recognition (CAER) is a complex and significant task that requires perceiving emotions from various contextual cues. Previous approaches primarily focus on designing sophisticated architectures to extract emotional cues from images. However, their knowledge is confined to specific training datasets and may reflect the subjective emotional biases of the annotators. Furthermore, acquiring large amounts of labeled data is often challenging in real-world applications. In this paper, we systematically explore the potential of leveraging Large Vision-Language Models (LVLMs) to empower the CAER task from three paradigms: 1) We fine-tune LVLMs on two CAER datasets, which is the most common way to transfer large models to downstream tasks. 2) We design zero-shot and few-shot patterns to evaluate the performance of LVLMs in scenarios with limited data or even completely unseen. In this case, a training-free framework is proposed to fully exploit the In-Context Learning (ICL) capabilities of LVLMs. Specifically, we develop an image similarity-based ranking algorithm to retrieve examples; subsequently, the instructions, retrieved examples, and the test example are combined to feed LVLMs to obtain the corresponding sentiment judgment. 3) To leverage the rich knowledge base of LVLMs, we incorporate Chain-of-Thought (CoT) into our framework to enhance the model's reasoning ability and provide interpretable results. Extensive experiments and analyses demonstrate that LVLMs achieve competitive performance in the CAER task across different paradigms. Notably, the superior performance in few-shot settings indicates the feasibility of LVLMs for accomplishing specific tasks without extensive training.

Dingkang Yang, Dongling Xiao, Jinjie Wei et al.

Despite their remarkable capabilities, Large Language Models (LLMs) are prone to generate responses that contradict verifiable facts, i.e., unfaithful hallucination content. Existing efforts generally focus on optimizing model parameters or editing semantic representations, which compromise the internal factual knowledge of target LLMs. In addition, hallucinations typically exhibit multifaceted patterns in downstream tasks, limiting the model's holistic performance across tasks. In this paper, we propose a Comparator-driven Decoding-Time (CDT) framework to alleviate the response hallucination. Firstly, we construct hallucinatory and truthful comparators with multi-task fine-tuning samples. In this case, we present an instruction prototype-guided mixture of experts strategy to enhance the ability of the corresponding comparators to capture different hallucination or truthfulness patterns in distinct task instructions. CDT constrains next-token predictions to factuality-robust distributions by contrasting the logit differences between the target LLMs and these comparators. Systematic experiments on multiple downstream tasks show that our framework can significantly improve the model performance and response factuality.

Xiao Zhao, Bo Chen, Mingyang Sun et al.

Vision-based 3D semantic scene completion (SSC) describes autonomous driving scenes through 3D volume representations. However, the occlusion of invisible voxels by scene surfaces poses challenges to current SSC methods in hallucinating refined 3D geometry. This paper proposes HybridOcc, a hybrid 3D volume query proposal method generated by Transformer framework and NeRF representation and refined in a coarse-to-fine SSC prediction framework. HybridOcc aggregates contextual features through the Transformer paradigm based on hybrid query proposals while combining it with NeRF representation to obtain depth supervision. The Transformer branch contains multiple scales and uses spatial cross-attention for 2D to 3D transformation. The newly designed NeRF branch implicitly infers scene occupancy through volume rendering, including visible and invisible voxels, and explicitly captures scene depth rather than generating RGB color. Furthermore, we present an innovative occupancy-aware ray sampling method to orient the SSC task instead of focusing on the scene surface, further improving the overall performance. Extensive experiments on nuScenes and SemanticKITTI datasets demonstrate the effectiveness of our HybridOcc on the SSC task.

Dingkang Yang, Kun Yang, Haopeng Kuang et al.

Understanding emotions from diverse contexts has received widespread attention in computer vision communities. The core philosophy of Context-Aware Emotion Recognition (CAER) is to provide valuable semantic cues for recognizing the emotions of target persons by leveraging rich contextual information. Current approaches invariably focus on designing sophisticated structures to extract perceptually critical representations from contexts. Nevertheless, a long-neglected dilemma is that a severe context bias in existing datasets results in an unbalanced distribution of emotional states among different contexts, causing biased visual representation learning. From a causal demystification perspective, the harmful bias is identified as a confounder that misleads existing models to learn spurious correlations based on likelihood estimation, limiting the models' performance. To address the issue, we embrace causal inference to disentangle the models from the impact of such bias, and formulate the causalities among variables in the CAER task via a customized causal graph. Subsequently, we present a Contextual Causal Intervention Module (CCIM) to de-confound the confounder, which is built upon backdoor adjustment theory to facilitate seeking approximate causal effects during model training. As a plug-and-play component, CCIM can easily integrate with existing approaches and bring significant improvements. Systematic experiments on three datasets demonstrate the effectiveness of our CCIM.

Mingyang Sun, Dingkang Yang, Dongliang Kou et al.

A human 3D avatar is one of the important elements in the metaverse, and the modeling effect directly affects people's visual experience. However, the human body has a complex topology and diverse details, so it is often expensive, time-consuming, and laborious to build a satisfactory model. Recent studies have proposed a novel method, implicit neural representation, which is a continuous representation method and can describe objects with arbitrary topology at arbitrary resolution. Researchers have applied implicit neural representation to human 3D avatar modeling and obtained more excellent results than traditional methods. This paper comprehensively reviews the application of implicit neural representation in human body modeling. First, we introduce three implicit representations of occupancy field, SDF, and NeRF, and make a classification of the literature investigated in this paper. Then the application of implicit modeling methods in the body, hand, and head are compared and analyzed respectively. Finally, we point out the shortcomings of current work and provide available suggestions for researchers.

Yuzheng Wang, Zhaoyu Chen, Dingkang Yang et al.

Adversarial Robustness Distillation (ARD) is a promising task to solve the issue of limited adversarial robustness of small capacity models while optimizing the expensive computational costs of Adversarial Training (AT). Despite the good robust performance, the existing ARD methods are still impractical to deploy in natural high-security scenes due to these methods rely entirely on original or publicly available data with a similar distribution. In fact, these data are almost always private, specific, and distinctive for scenes that require high robustness. To tackle these issues, we propose a challenging but significant task called Data-Free Adversarial Robustness Distillation (DFARD), which aims to train small, easily deployable, robust models without relying on data. We demonstrate that the challenge lies in the lower upper bound of knowledge transfer information, making it crucial to mining and transferring knowledge more efficiently. Inspired by human education, we design a plug-and-play Interactive Temperature Adjustment (ITA) strategy to improve the efficiency of knowledge transfer and propose an Adaptive Generator Balance (AGB) module to retain more data information. Our method uses adaptive hyperparameters to avoid a large number of parameter tuning, which significantly outperforms the combination of existing techniques. Meanwhile, our method achieves stable and reliable performance on multiple benchmarks.

Xiao Zhao, Xukun Zhang, Dingkang Yang et al.

Accurate and robust multimodal multi-task perception is crucial for modern autonomous driving systems. However, current multimodal perception research follows independent paradigms designed for specific perception tasks, leading to a lack of complementary learning among tasks and decreased performance in multi-task learning (MTL) due to joint training. In this paper, we propose MaskBEV, a masked attention-based MTL paradigm that unifies 3D object detection and bird's eye view (BEV) map segmentation. MaskBEV introduces a task-agnostic Transformer decoder to process these diverse tasks, enabling MTL to be completed in a unified decoder without requiring additional design of specific task heads. To fully exploit the complementary information between BEV map segmentation and 3D object detection tasks in BEV space, we propose spatial modulation and scene-level context aggregation strategies. These strategies consider the inherent dependencies between BEV segmentation and 3D detection, naturally boosting MTL performance. Extensive experiments on nuScenes dataset show that compared with previous state-of-the-art MTL methods, MaskBEV achieves 1.3 NDS improvement in 3D object detection and 2.7 mIoU improvement in BEV map segmentation, while also demonstrating slightly leading inference speed.

Shunli Wang, Qing Yu, Shuaibing Wang et al.

The fine-grained medical action analysis task has received considerable attention from pattern recognition communities recently, but it faces the problems of data and algorithm shortage. Cardiopulmonary Resuscitation (CPR) is an essential skill in emergency treatment. Currently, the assessment of CPR skills mainly depends on dummies and trainers, leading to high training costs and low efficiency. For the first time, this paper constructs a vision-based system to complete error action recognition and skill assessment in CPR. Specifically, we define 13 types of single-error actions and 74 types of composite error actions during external cardiac compression and then develop a video dataset named CPR-Coach. By taking the CPR-Coach as a benchmark, this paper thoroughly investigates and compares the performance of existing action recognition models based on different data modalities. To solve the unavoidable Single-class Training & Multi-class Testing problem, we propose a humancognition-inspired framework named ImagineNet to improve the model's multierror recognition performance under restricted supervision. Extensive experiments verify the effectiveness of the framework. We hope this work could advance research toward fine-grained medical action analysis and skill assessment. The CPR-Coach dataset and the code of ImagineNet are publicly available on Github.

Yuzheng Wang, Zhaoyu Chen, Dingkang Yang et al.

Novel Class Discovery (NCD) aims to discover unknown and novel classes in an unlabeled set by leveraging knowledge already learned about known classes. Existing works focus on instance-level or class-level knowledge representation and build a shared representation space to achieve performance improvements. However, a long-neglected issue is the potential imbalanced number of samples from known and novel classes, pushing the model towards dominant classes. Therefore, these methods suffer from a challenging trade-off between reviewing known classes and discovering novel classes. Based on this observation, we propose a Self-Cooperation Knowledge Distillation (SCKD) method to utilize each training sample (whether known or novel, labeled or unlabeled) for both review and discovery. Specifically, the model's feature representations of known and novel classes are used to construct two disjoint representation spaces. Through spatial mutual information, we design a self-cooperation learning to encourage model learning from the two feature representation spaces from itself. Extensive experiments on six datasets demonstrate that our method can achieve significant performance improvements, achieving state-of-the-art performance.

Wei Xue, Mingcheng Li, Xuecheng Wu et al.

Vision-and-Language Navigation (VLN) requires agents to accurately perceive complex visual environments and reason over navigation instructions and histories. However, existing methods passively process redundant visual inputs and treat all historical contexts indiscriminately, resulting in inefficient perception and unfocused reasoning. To address these challenges, we propose \textbf{ProFocus}, a training-free progressive framework that unifies \underline{Pro}active Perception and \underline{Focus}ed Reasoning through collaboration between large language models (LLMs) and vision-language models (VLMs). For proactive perception, ProFocus transforms panoramic observations into structured ego-centric semantic maps, enabling the orchestration agent to identify missing visual information needed for reliable decision-making, and to generate targeted visual queries with corresponding focus regions that guide the perception agent to acquire the required observations. For focused reasoning, we propose Branch-Diverse Monte Carlo Tree Search (BD-MCTS) to identify top-$k$ high-value waypoints from extensive historical candidates. The decision agent focuses reasoning on the historical contexts associated with these waypoints, rather than considering all historical waypoints equally. Extensive experiments validate the effectiveness of ProFocus, achieving state-of-the-art performance among zero-shot methods on R2R and REVERIE benchmarks.

Zizhi Chen, Yizhen Gao, Minghao Han et al.

Multimodal biomedical Vision-Language Models (VLMs) exhibit immense potential in the field of Continual Learning (CL). However, they confront a core dilemma: how to preserve fine-grained intra-modality features while bridging the significant domain gap across different modalities. To address this challenge, we propose a comprehensive framework. Leveraging our 18-million multimodal and comprehensive medical retrieval database derived from PubMed scientific papers, we pioneer the integration of Retrieval-Augmented Generation (RAG) into CL. Specifically, we employ a multi-modal, multi-layer RAG system that provides real-time guidance for model fine-tuning through dynamic, on-demand knowledge retrieval. Building upon this, we introduce a dynamic knowledge distillation framework. This framework precisely resolves the aforementioned core dilemma by dynamically modulating the importance of the parameter space, the granularity of the distilled knowledge, and the data distribution of the reference dataset in accordance with the required level of detail. To thoroughly validate the clinical value of our strategy, we have designed a more rigorous \textbf{M}edical Generalist Task Incremental Learning (MGTIL) benchmark. This benchmark is engineered to simultaneously evaluate the model's capacity for adaptation to significant domain shifts, retention of subtle intra-domain features, and real-time learning of novel and complex medical tasks. Extensive experimental results demonstrate that our proposed method achieves state-of-the-art (SOTA) performance across all metrics. The code is provided in the supplementary materials.

Biaoyu Ren, Qingsheng Wang, Cun Xu et al.

Referring Remote Sensing Image Segmentation (RRSIS) is a situated, task-driven cross-modal task related to the embodied perception paradigm, requiring models to align visual-spatial features with linguistic intentions for precise target perception. Recent research has focused on refining the granularity of textual features and optimizing image-text feature fusion to better guide target feature representations. However, insufficient descriptive granularity and sensitivity to semantic shifts can cause bottlenecks in cross-modal feature fusion. To address these issues, we propose the Image-Conditioned Instance Prompt Network (ICIPNet) with Bilateral Information Fusion, which is designed to alleviate bottlenecks in cross-modal feature fusion. ICIPNet introduces an Image-Conditioned Instance Prompt (ICIP) module to generate self-adaptive visual and semantic representations without external knowledge. The Bilateral Information Fusion (BIF) module enhances feature fusion along the token and channel dimensions. Experiments demonstrate that the proposed ICIPNet outperforms existing RRSIS models.

Minghao Han, Dingkang Yang, Yue Jiang et al.

Physical understanding remains brittle in omni-modal models because key physical attributes are visually ambiguous and sparsely represented in web-scale data. We present OmniFysics, a compact omni-modal model that unifies understanding across images, audio, video, and text, with integrated speech and image generation. To inject explicit physical knowledge, we build a physical data engine with two components. FysicsAny produces physics-grounded instruction--image supervision by mapping salient objects to verified physical attributes through hierarchical retrieval over a curated prototype database, followed by physics-law--constrained verification and caption rewriting. FysicsOmniCap distills web videos via audio--visual consistency filtering to generate high-fidelity video--instruction pairs emphasizing cross-modal physical cues. We train OmniFysics with staged multimodal alignment and instruction tuning, adopt latent-space flow matching for text-to-image generation, and use an intent router to activate generation only when needed. Experiments show competitive performance on standard multimodal benchmarks and improved results on physics-oriented evaluations.

Dingkang Yang, Mingcheng Li, Xuecheng Wu et al.

Multimodal Sentiment Analysis (MSA) aims to predict sentiment from language, acoustic, and visual data in videos. However, imbalanced unimodal performance often leads to suboptimal fused representations. Existing approaches typically adopt fixed primary modality strategies to maximize dominant modality advantages, yet fail to adapt to dynamic variations in modality importance across different samples. Moreover, non-language modalities suffer from sequential redundancy and noise, degrading model performance when they serve as primary inputs. To address these issues, this paper proposes a modality optimization and dynamic primary modality selection framework (MODS). First, a Graph-based Dynamic Sequence Compressor (GDC) is constructed, which employs capsule networks and graph convolution to reduce sequential redundancy in acoustic/visual modalities. Then, we develop a sample-adaptive Primary Modality Selector (MSelector) for dynamic dominance determination. Finally, a Primary-modality-Centric Cross-Attention (PCCA) module is designed to enhance dominant modalities while facilitating cross-modal interaction. Extensive experiments on four benchmark datasets demonstrate that MODS outperforms state-of-the-art methods, achieving superior performance by effectively balancing modality contributions and eliminating redundant noise.

Shuaibing Wang, Shunli Wang, Mingcheng Li et al.

Temporal Action Segmentation (TAS) is an essential task in video analysis, aiming to segment and classify continuous frames into distinct action segments. However, the ambiguous boundaries between actions pose a significant challenge for high-precision segmentation. Recent advances in diffusion models have demonstrated substantial success in TAS tasks due to their stable training process and high-quality generation capabilities. However, the heavy sampling steps required by diffusion models pose a substantial computational burden, limiting their practicality in real-time applications. Additionally, most related works utilize Transformer-based encoder architectures. Although these architectures excel at capturing long-range dependencies, they incur high computational costs and face feature-smoothing issues when processing long video sequences. To address these challenges, we propose EffiDiffAct, an efficient and high-performance TAS algorithm. Specifically, we develop a lightweight temporal feature encoder that reduces computational overhead and mitigates the rank collapse phenomenon associated with traditional self-attention mechanisms. Furthermore, we introduce an adaptive skip strategy that allows for dynamic adjustment of timestep lengths based on computed similarity metrics during inference, thereby further enhancing computational efficiency. Comprehensive experiments on the 50Salads, Breakfast, and GTEA datasets demonstrated the effectiveness of the proposed algorithm.

Hanshen Zhu, Yuliang Liu, Xuecheng Wu et al.

Visual Text Rendering (VTR) remains a critical challenge in text-to-image generation, where even advanced models frequently produce text with structural anomalies such as distortion, blurriness, and misalignment. However, we find that leading MLLMs and specialist OCR models largely fail to perceive these structural anomalies, creating a critical bottleneck for both VTR evaluation and RL-based optimization. As a result, even state-of-the-art generators (e.g., Seedream4.0, Qwen-Image) still struggle to render structurally faithful text. To address this, we propose TextPecker, a plug-and-play structural anomaly perceptive RL strategy that mitigates noisy reward signals and works with any textto-image generator. To enable this capability, we construct a recognition dataset with character-level structural-anomaly annotations and develop a stroke-editing synthesis engine to expand structural-error coverage. Experiments show that TextPecker consistently improves diverse text-to-image models; even on the well-optimized Qwen-Image, it significantly yields average gains of 4% in structural fidelity and 8.7% in semantic alignment for Chinese text rendering, establishing a new state-of-the-art in high-fidelity VTR. Our work fills a gap in VTR optimization, providing a foundational step towards reliable and structural faithful visual text generation.

Jiawei Chen, Yue Jiang, Dingkang Yang et al.

While Large Language Models (LLMs) excel in world knowledge understanding, adapting them to specific subfields requires precise adjustments. Due to the model's vast scale, traditional global fine-tuning methods for large models can be computationally expensive and impact generalization. To address this challenge, a range of innovative Parameters-Efficient Fine-Tuning (PEFT) methods have emerged and achieved remarkable success in both LLMs and Large Vision-Language Models (LVLMs). In the medical domain, fine-tuning a medical Vision-Language Pretrained (VLP) model is essential for adapting it to specific tasks. Can the fine-tuning methods for large models be transferred to the medical field to enhance transfer learning efficiency? In this paper, we delve into the fine-tuning methods of LLMs and conduct extensive experiments to investigate the impact of fine-tuning methods for large models on the existing multimodal model in the medical domain from the training data level and the model structure level. We show the different impacts of fine-tuning methods for large models on medical VLMs and develop the most efficient ways to fine-tune medical VLP models. We hope this research can guide medical domain researchers in optimizing VLMs' training costs, fostering the broader application of VLMs in healthcare fields. The code and dataset have been released at https://github.com/TIMMY-CHAN/MILE.

Shuaibing Wang, Shunli Wang, Dingkang Yang et al.

We propose a robust and accurate method for reconstructing 3D hand mesh from monocular images. This is a very challenging problem, as hands are often severely occluded by objects. Previous works often have disregarded 2D hand pose information, which contains hand prior knowledge that is strongly correlated with occluded regions. Thus, in this work, we propose a novel 3D hand mesh reconstruction network HandGCAT, that can fully exploit hand prior as compensation information to enhance occluded region features. Specifically, we designed the Knowledge-Guided Graph Convolution (KGC) module and the Cross-Attention Transformer (CAT) module. KGC extracts hand prior information from 2D hand pose by graph convolution. CAT fuses hand prior into occluded regions by considering their high correlation. Extensive experiments on popular datasets with challenging hand-object occlusions, such as HO3D v2, HO3D v3, and DexYCB demonstrate that our HandGCAT reaches state-of-the-art performance. The code is available at https://github.com/heartStrive/HandGCAT.

Hao Feng, Wei Shi, Ke Zhang et al.

Document parsing has garnered widespread attention as vision-language models (VLMs) advance OCR capabilities. However, the field remains fragmented across dozens of specialized models with varying strengths, forcing users to navigate complex model selection and limiting system scalability. Moreover, existing two-stage approaches depend on axis-aligned bounding boxes for layout detection, failing to handle distorted or photographed documents effectively. To this end, we present Dolphin-v2, a two-stage document image parsing model that substantially improves upon the original Dolphin. In the first stage, Dolphin-v2 jointly performs document type classification (digital-born versus photographed) alongside layout analysis. For digital-born documents, it conducts finer-grained element detection with reading order prediction. In the second stage, we employ a hybrid parsing strategy: photographed documents are parsed holistically as complete pages to handle geometric distortions, while digital-born documents undergo element-wise parallel parsing guided by the detected layout anchors, enabling efficient content extraction. Compared with the original Dolphin, Dolphin-v2 introduces several crucial enhancements: (1) robust parsing of photographed documents via holistic page-level understanding, (2) finer-grained element detection (21 categories) with semantic attribute extraction such as author information and document metadata, and (3) code block recognition with indentation preservation, which existing systems typically lack. Comprehensive evaluations are conducted on DocPTBench, OmniDocBench, and our self-constructed RealDoc-160 benchmark. The results demonstrate substantial improvements: +14.78 points overall on the challenging OmniDocBench and 91% error reduction on photographed documents, while maintaining efficient inference through parallel processing.

Minghao Han, Dingkang Yang, Jiabei Cheng et al.

Recent advancements in multimodal pre-training models have significantly advanced computational pathology. However, current approaches predominantly rely on visual-language models, which may impose limitations from a molecular perspective and lead to performance bottlenecks. Here, we introduce a Unified Molecule-enhanced Pathology Image REpresentationn Learning framework (UMPIRE). UMPIRE aims to leverage complementary information from gene expression profiles to guide the multimodal pre-training, enhancing the molecular awareness of pathology image representation learning. We demonstrate that this molecular perspective provides a robust, task-agnostic training signal for learning pathology image embeddings. Due to the scarcity of paired data, approximately 4 million entries of spatial transcriptomics gene expression were collected to train the gene encoder. By leveraging powerful pre-trained encoders, UMPIRE aligns the encoders across over 697K pathology image-gene expression pairs. The performance of UMPIRE is demonstrated across various molecular-related downstream tasks, including gene expression prediction, spot classification, and mutation state prediction in whole slide images. Our findings highlight the effectiveness of multimodal data integration and open new avenues for exploring computational pathology enhanced by molecular perspectives. The code and pre-trained weights are available at https://github.com/Hanminghao/UMPIRE.

Weijie Yin, Dingkang Yang, Hongyuan Dong et al.

Vision Transformers (ViTs) are essential as foundation backbones in establishing the visual comprehension capabilities of Multimodal Large Language Models (MLLMs). Although most ViTs achieve impressive performance through image-text pair-based contrastive learning or self-supervised mechanisms, they struggle to engage in connector-based co-training directly with LLMs due to potential parameter initialization conflicts and modality semantic gaps. To address the above challenges, this paper proposes SAILViT, a gradual feature learning-enhanced ViT for facilitating MLLMs to break through performance bottlenecks in complex multimodal interactions. SAILViT achieves coarse-to-fine-grained feature alignment and world knowledge infusion with gradual feature refinement, which better serves target training demands. We perform thorough empirical analyses to confirm the powerful robustness and generalizability of SAILViT across different dimensions, including parameter sizes, model architectures, training strategies, and data scales. Equipped with SAILViT, existing MLLMs show significant and consistent performance improvements on the OpenCompass benchmark across extensive downstream tasks. SAILViT series models are released at https://huggingface.co/BytedanceDouyinContent.

Keliang Liu, Dingkang Yang, Ziyun Qian et al.

In recent years, training methods centered on Reinforcement Learning (RL) have markedly enhanced the reasoning and alignment performance of Large Language Models (LLMs), particularly in understanding human intents, following user instructions, and bolstering inferential strength. Although existing surveys offer overviews of RL augmented LLMs, their scope is often limited, failing to provide a comprehensive summary of how RL operates across the full lifecycle of LLMs. We systematically review the theoretical and practical advancements whereby RL empowers LLMs, especially Reinforcement Learning with Verifiable Rewards (RLVR). First, we briefly introduce the basic theory of RL. Second, we thoroughly detail application strategies for RL across various phases of the LLM lifecycle, including pre-training, alignment fine-tuning, and reinforced reasoning. In particular, we emphasize that RL methods in the reinforced reasoning phase serve as a pivotal driving force for advancing model reasoning to its limits. Next, we collate existing datasets and evaluation benchmarks currently used for RL fine-tuning, spanning human-annotated datasets, AI-assisted preference data, and program-verification-style corpora. Subsequently, we review the mainstream open-source tools and training frameworks available, providing clear practical references for subsequent research. Finally, we analyse the future challenges and trends in the field of RL-enhanced LLMs. This survey aims to present researchers and practitioners with the latest developments and frontier trends at the intersection of RL and LLMs, with the goal of fostering the evolution of LLMs that are more intelligent, generalizable, and secure.

Weijie Yin, Yongjie Ye, Fangxun Shu et al.

We introduce SAIL-VL2, an open-suite vision-language foundation model (LVM) for comprehensive multimodal understanding and reasoning. As the successor to SAIL-VL, SAIL-VL2 achieves state-of-the-art performance at the 2B and 8B parameter scales across diverse image and video benchmarks, demonstrating strong capabilities from fine-grained perception to complex reasoning. Its effectiveness is driven by three core innovations. First, a large-scale data curation pipeline with scoring and filtering strategies enhances both quality and distribution across captioning, OCR, QA, and video data, improving training efficiency. Second, a progressive training framework begins with a powerful pre-trained vision encoder (SAIL-ViT), advances through multimodal pre-training, and culminates in a thinking-fusion SFT-RL hybrid paradigm that systematically strengthens model capabilities. Third, architectural advances extend beyond dense LLMs to efficient sparse Mixture-of-Experts (MoE) designs. With these contributions, SAIL-VL2 demonstrates competitive performance across 106 datasets and achieves state-of-the-art results on challenging reasoning benchmarks such as MMMU and MathVista. Furthermore, on the OpenCompass leaderboard, SAIL-VL2-2B ranks first among officially released open-source models under the 4B parameter scale, while serving as an efficient and extensible foundation for the open-source multimodal community.

Hao Yang, Hongyuan Lu, Dingkang Yang et al.

Instant-messaging human social chat typically progresses through a sequence of short messages. Existing step-by-step AI chatting systems typically split a one-shot generation into multiple messages and send them sequentially, but they lack an active waiting mechanism and exhibit unnatural message pacing. In order to address these issues, we propose Stephanie2, a novel next-generation step-wise decision-making dialogue agent. With active waiting and message-pace adaptation, Stephanie2 explicitly decides at each step whether to send or wait, and models latency as the sum of thinking time and typing time to achieve more natural pacing. We further introduce a time-window-based dual-agent dialogue system to generate pseudo dialogue histories for human and automatic evaluations. Experiments show that Stephanie2 clearly outperforms Stephanie1 on metrics such as naturalness and engagement, and achieves a higher pass rate on human evaluation with the role identification Turing test.

Shi Chen, Xuecheng Wu, Heli Sun et al.

Affective Image Manipulation (AIM) aims to evoke specific emotions through targeted editing. Current image editing benchmarks primarily focus on object-level modifications in general scenarios, lacking the fine-grained granularity to capture affective dimensions. To bridge this gap, we introduce the first benchmark designed for AIM termed AIM-Bench. This benchmark is built upon a dual-path affective modeling scheme that integrates the Mikels emotion taxonomy with the Valence-Arousal-Dominance framework, enabling high-level semantic and fine-grained continuous manipulation. Through a hierarchical human-in-the-loop workflow, we finally curate 800 high-quality samples covering 8 emotional categories and 5 editing types. To effectively assess performance, we also design a composite evaluation suite combining rule-based and model-based metrics to holistically assess instruction consistency, aesthetics, and emotional expressiveness. Extensive evaluations reveal that current editing models face significant challenges, most notably a prevalent positivity bias, which stemming from inherent imbalances in training data distribution. To tackle this, we propose a scalable data engine utilizing an inverse repainting strategy to construct AIM-40k, a balanced instruction-tuning dataset comprising 40k samples. Concretely, we enhance raw affective images via generative redrawing to establish high-fidelity ground truths, and synthesize input images with divergent emotions and paired precise instructions. Fine-tuning a baseline model on AIM-40k yields a 9.15% relative improvement in overall performance, demonstrating the effectiveness of our AIM-40k. Our data and related code will be made open soon.

Yue Jiang, Jichu Li, Yang Liu et al.

We introduce DanmakuTPPBench, a comprehensive benchmark designed to advance multi-modal Temporal Point Process (TPP) modeling in the era of Large Language Models (LLMs). While TPPs have been widely studied for modeling temporal event sequences, existing datasets are predominantly unimodal, hindering progress in models that require joint reasoning over temporal, textual, and visual information. To address this gap, DanmakuTPPBench comprises two complementary components: (1) DanmakuTPP-Events, a novel dataset derived from the Bilibili video platform, where user-generated bullet comments (Danmaku) naturally form multi-modal events annotated with precise timestamps, rich textual content, and corresponding video frames; (2) DanmakuTPP-QA, a challenging question-answering dataset constructed via a novel multi-agent pipeline powered by state-of-the-art LLMs and multi-modal LLMs (MLLMs), targeting complex temporal-textual-visual reasoning. We conduct extensive evaluations using both classical TPP models and recent MLLMs, revealing significant performance gaps and limitations in current methods' ability to model multi-modal event dynamics. Our benchmark establishes strong baselines and calls for further integration of TPP modeling into the multi-modal language modeling landscape. Project page: https://github.com/FRENKIE-CHIANG/DanmakuTPPBench

Minghao Han, Xukun Zhang, Dingkang Yang et al.

Survival prediction is a complex ordinal regression task that aims to predict the survival coefficient ranking among a cohort of patients, typically achieved by analyzing patients' whole slide images. Existing deep learning approaches mainly adopt multiple instance learning or graph neural networks under weak supervision. Most of them are unable to uncover the diverse interactions between different types of biological entities(\textit{e.g.}, cell cluster and tissue block) across multiple scales, while such interactions are crucial for patient survival prediction. In light of this, we propose a novel multi-scale heterogeneity-aware hypergraph representation framework. Specifically, our framework first constructs a multi-scale heterogeneity-aware hypergraph and assigns each node with its biological entity type. It then mines diverse interactions between nodes on the graph structure to obtain a global representation. Experimental results demonstrate that our method outperforms state-of-the-art approaches on three benchmark datasets. Code is publicly available at \href{https://github.com/Hanminghao/H2GT}{https://github.com/Hanminghao/H2GT}.

Minghao Han, Dingkang Yang, Linhao Qu et al.

Recent years have witnessed remarkable progress in multimodal learning within computational pathology. Existing models primarily rely on vision and language modalities; however, language alone lacks molecular specificity and offers limited pathological supervision, leading to representational bottlenecks. In this paper, we propose STAMP, a Spatial Transcriptomics-Augmented Multimodal Pathology representation learning framework that integrates spatially-resolved gene expression profiles to enable molecule-guided joint embedding of pathology images and transcriptomic data. Our study shows that self-supervised, gene-guided training provides a robust and task-agnostic signal for learning pathology image representations. Incorporating spatial context and multi-scale information further enhances model performance and generalizability. To support this, we constructed SpaVis-6M, the largest Visium-based spatial transcriptomics dataset to date, and trained a spatially-aware gene encoder on this resource. Leveraging hierarchical multi-scale contrastive alignment and cross-scale patch localization mechanisms, STAMP effectively aligns spatial transcriptomics with pathology images, capturing spatial structure and molecular variation. We validate STAMP across six datasets and four downstream tasks, where it consistently achieves strong performance. These results highlight the value and necessity of integrating spatially resolved molecular supervision for advancing multimodal learning in computational pathology. The code is included in the supplementary materials. The pretrained weights and SpaVis-6M are available at: https://github.com/Hanminghao/STAMP.

Yuxiang Nie, Han Wang, Yongjie Ye et al.

This paper introduces ChineseVideoBench, a pioneering benchmark specifically designed for evaluating Multimodal Large Language Models (MLLMs) in Chinese Video Question Answering. The growing demand for sophisticated video analysis capabilities highlights the critical need for comprehensive, culturally-aware evaluation frameworks. ChineseVideoBench addresses this gap by providing a robust dataset and tailored evaluation metrics, enabling rigorous assessment of state-of-the-art MLLMs on complex Chinese video content. Specifically, ChineseVideoBench comprises 8 main classes and 12 sub-classes, encompassing tasks that demand both deep video understanding and nuanced Chinese linguistic and cultural awareness. Our empirical evaluations reveal that ChineseVideoBench presents a significant challenge to current MLLMs. Among the models assessed, Gemini 2.5 Pro achieves the highest performance with an overall score of 77.9%, while InternVL-38B emerges as the most competitive open-source model.