Yuming Jiang

Jun Cen, Siteng Huang, Yuqian Yuan et al. · pku

We introduce RynnVLA-002, a unified Vision-Language-Action (VLA) and world model. The world model leverages action and visual inputs to predict future image states, learning the underlying physics of the environment to refine action generation. Conversely, the VLA model produces subsequent actions from image observations, enhancing visual understanding and supporting the world model's image generation. The unified framework of RynnVLA-002 enables joint learning of environmental dynamics and action planning. Our experiments show that RynnVLA-002 surpasses individual VLA and world models, demonstrating their mutual enhancement. We evaluate RynnVLA-002 in both simulation and real-world robot tasks. RynnVLA-002 achieves 97.4% success rate on the LIBERO simulation benchmark without pretraining, while in real-world LeRobot experiments, its integrated world model boosts the overall success rate by 50%.

Ziqi Huang, Yinan He, Jiashuo Yu et al.

Video generation has witnessed significant advancements, yet evaluating these models remains a challenge. A comprehensive evaluation benchmark for video generation is indispensable for two reasons: 1) Existing metrics do not fully align with human perceptions; 2) An ideal evaluation system should provide insights to inform future developments of video generation. To this end, we present VBench, a comprehensive benchmark suite that dissects "video generation quality" into specific, hierarchical, and disentangled dimensions, each with tailored prompts and evaluation methods. VBench has three appealing properties: 1) Comprehensive Dimensions: VBench comprises 16 dimensions in video generation (e.g., subject identity inconsistency, motion smoothness, temporal flickering, and spatial relationship, etc). The evaluation metrics with fine-grained levels reveal individual models' strengths and weaknesses. 2) Human Alignment: We also provide a dataset of human preference annotations to validate our benchmarks' alignment with human perception, for each evaluation dimension respectively. 3) Valuable Insights: We look into current models' ability across various evaluation dimensions, and various content types. We also investigate the gaps between video and image generation models. We will open-source VBench, including all prompts, evaluation methods, generated videos, and human preference annotations, and also include more video generation models in VBench to drive forward the field of video generation.

Yaohui Wang, Xinyuan Chen, Xin Ma et al.

This work aims to learn a high-quality text-to-video (T2V) generative model by leveraging a pre-trained text-to-image (T2I) model as a basis. It is a highly desirable yet challenging task to simultaneously a) accomplish the synthesis of visually realistic and temporally coherent videos while b) preserving the strong creative generation nature of the pre-trained T2I model. To this end, we propose LaVie, an integrated video generation framework that operates on cascaded video latent diffusion models, comprising a base T2V model, a temporal interpolation model, and a video super-resolution model. Our key insights are two-fold: 1) We reveal that the incorporation of simple temporal self-attentions, coupled with rotary positional encoding, adequately captures the temporal correlations inherent in video data. 2) Additionally, we validate that the process of joint image-video fine-tuning plays a pivotal role in producing high-quality and creative outcomes. To enhance the performance of LaVie, we contribute a comprehensive and diverse video dataset named Vimeo25M, consisting of 25 million text-video pairs that prioritize quality, diversity, and aesthetic appeal. Extensive experiments demonstrate that LaVie achieves state-of-the-art performance both quantitatively and qualitatively. Furthermore, we showcase the versatility of pre-trained LaVie models in various long video generation and personalized video synthesis applications.

Jianglin Fu, Shikai Li, Yuming Jiang et al.

Unconditional human image generation is an important task in vision and graphics, which enables various applications in the creative industry. Existing studies in this field mainly focus on "network engineering" such as designing new components and objective functions. This work takes a data-centric perspective and investigates multiple critical aspects in "data engineering", which we believe would complement the current practice. To facilitate a comprehensive study, we collect and annotate a large-scale human image dataset with over 230K samples capturing diverse poses and textures. Equipped with this large dataset, we rigorously investigate three essential factors in data engineering for StyleGAN-based human generation, namely data size, data distribution, and data alignment. Extensive experiments reveal several valuable observations w.r.t. these aspects: 1) Large-scale data, more than 40K images, are needed to train a high-fidelity unconditional human generation model with vanilla StyleGAN. 2) A balanced training set helps improve the generation quality with rare face poses compared to the long-tailed counterpart, whereas simply balancing the clothing texture distribution does not effectively bring an improvement. 3) Human GAN models with body centers for alignment outperform models trained using face centers or pelvis points as alignment anchors. In addition, a model zoo and human editing applications are demonstrated to facilitate future research in the community.

Yuming Jiang, Shuai Yang, Tong Liang Koh et al.

Text-driven content creation has evolved to be a transformative technique that revolutionizes creativity. Here we study the task of text-driven human video generation, where a video sequence is synthesized from texts describing the appearance and motions of a target performer. Compared to general text-driven video generation, human-centric video generation requires maintaining the appearance of synthesized human while performing complex motions. In this work, we present Text2Performer to generate vivid human videos with articulated motions from texts. Text2Performer has two novel designs: 1) decomposed human representation and 2) diffusion-based motion sampler. First, we decompose the VQVAE latent space into human appearance and pose representation in an unsupervised manner by utilizing the nature of human videos. In this way, the appearance is well maintained along the generated frames. Then, we propose continuous VQ-diffuser to sample a sequence of pose embeddings. Unlike existing VQ-based methods that operate in the discrete space, continuous VQ-diffuser directly outputs the continuous pose embeddings for better motion modeling. Finally, motion-aware masking strategy is designed to mask the pose embeddings spatial-temporally to enhance the temporal coherence. Moreover, to facilitate the task of text-driven human video generation, we contribute a Fashion-Text2Video dataset with manually annotated action labels and text descriptions. Extensive experiments demonstrate that Text2Performer generates high-quality human videos (up to 512x256 resolution) with diverse appearances and flexible motions.

Yuming Jiang, Shuai Yang, Haonan Qiu et al.

Generating high-quality and diverse human images is an important yet challenging task in vision and graphics. However, existing generative models often fall short under the high diversity of clothing shapes and textures. Furthermore, the generation process is even desired to be intuitively controllable for layman users. In this work, we present a text-driven controllable framework, Text2Human, for a high-quality and diverse human generation. We synthesize full-body human images starting from a given human pose with two dedicated steps. 1) With some texts describing the shapes of clothes, the given human pose is first translated to a human parsing map. 2) The final human image is then generated by providing the system with more attributes about the textures of clothes. Specifically, to model the diversity of clothing textures, we build a hierarchical texture-aware codebook that stores multi-scale neural representations for each type of texture. The codebook at the coarse level includes the structural representations of textures, while the codebook at the fine level focuses on the details of textures. To make use of the learned hierarchical codebook to synthesize desired images, a diffusion-based transformer sampler with mixture of experts is firstly employed to sample indices from the coarsest level of the codebook, which then is used to predict the indices of the codebook at finer levels. The predicted indices at different levels are translated to human images by the decoder learned accompanied with hierarchical codebooks. The use of mixture-of-experts allows for the generated image conditioned on the fine-grained text input. The prediction for finer level indices refines the quality of clothing textures. Extensive quantitative and qualitative evaluations demonstrate that our proposed framework can generate more diverse and realistic human images compared to state-of-the-art methods.

Haonan Qiu, Zhaoxi Chen, Yuming Jiang et al.

Recent years have witnessed great progress in creating vivid audio-driven portraits from monocular videos. However, how to seamlessly adapt the created video avatars to other scenarios with different backgrounds and lighting conditions remains unsolved. On the other hand, existing relighting studies mostly rely on dynamically lighted or multi-view data, which are too expensive for creating video portraits. To bridge this gap, we propose ReliTalk, a novel framework for relightable audio-driven talking portrait generation from monocular videos. Our key insight is to decompose the portrait's reflectance from implicitly learned audio-driven facial normals and images. Specifically, we involve 3D facial priors derived from audio features to predict delicate normal maps through implicit functions. These initially predicted normals then take a crucial part in reflectance decomposition by dynamically estimating the lighting condition of the given video. Moreover, the stereoscopic face representation is refined using the identity-consistent loss under simulated multiple lighting conditions, addressing the ill-posed problem caused by limited views available from a single monocular video. Extensive experiments validate the superiority of our proposed framework on both real and synthetic datasets. Our code is released in https://github.com/arthur-qiu/ReliTalk.

Jianxiong Zhang, Bing Guo, Yuming Jiang et al.

Large reasoning models (LRMs) often generate long, seemingly coherent reasoning traces yet still produce incorrect answers, making hallucination detection challenging. Although trajectories contain useful signals, directly using trace text or vanilla hidden states for detection is brittle: traces vary in form and detectors can overfit to superficial patterns rather than answer validity. We introduce Answer-agreement Representation Shaping (ARS), which learns detection-friendly trace-conditioned representations by explicitly encoding answer stability. ARS generates counterfactual answers through small latent interventions, specifically, perturbing the trace-boundary embedding, and labels each perturbation by whether the resulting answer agrees with the original. It then learns representations that bring answer-agreeing states together and separate answer-disagreeing ones, exposing latent instability indicative of hallucination risk. The shaped embeddings are plug-and-play with existing embedding-based detectors and require no human annotations during training. Experiments demonstrate that ARS consistently improves detection and achieves substantial gains over strong baselines. Code is available at: https://github.com/radiolab-ntu/ars_icml2026.

Yuming Jiang, Kelvin C. K. Chan, Xintao Wang et al.

Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image or video by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g., scale and rotation) and the resolution gap (e.g., HR and LR). To tackle these challenges, we propose C2-Matching in this work, which performs explicit robust matching crossing transformation and resolution. 1) To bridge the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) To address the resolution gap, we adopt teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue between input images and reference images. In addition, to faithfully evaluate the performance of Reference-based Image Super-Resolution under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. We also extend C2-Matching to Reference-based Video Super-Resolution task, where an image taken in a similar scene serves as the HR reference image. Extensive experiments demonstrate that our proposed C2-Matching significantly outperforms state of the arts on the standard CUFED5 benchmark and also boosts the performance of video SR by incorporating the C2-Matching component into Video SR pipelines.

Yanyan Huang, Weiqin Zhao, Shujun Wang et al.

Whole slide image (WSI) analysis has become increasingly important in the medical imaging community, enabling automated and objective diagnosis, prognosis, and therapeutic-response prediction. However, in clinical practice, the ever-evolving environment hamper the utility of WSI analysis models. In this paper, we propose the FIRST continual learning framework for WSI analysis, named ConSlide, to tackle the challenges of enormous image size, utilization of hierarchical structure, and catastrophic forgetting by progressive model updating on multiple sequential datasets. Our framework contains three key components. The Hierarchical Interaction Transformer (HIT) is proposed to model and utilize the hierarchical structural knowledge of WSI. The Breakup-Reorganize (BuRo) rehearsal method is developed for WSI data replay with efficient region storing buffer and WSI reorganizing operation. The asynchronous updating mechanism is devised to encourage the network to learn generic and specific knowledge respectively during the replay stage, based on a nested cross-scale similarity learning (CSSL) module. We evaluated the proposed ConSlide on four public WSI datasets from TCGA projects. It performs best over other state-of-the-art methods with a fair WSI-based continual learning setting and achieves a better trade-off of the overall performance and forgetting on previous task

Ziqi Huang, Kelvin C. K. Chan, Yuming Jiang et al.

Diffusion models arise as a powerful generative tool recently. Despite the great progress, existing diffusion models mainly focus on uni-modal control, i.e., the diffusion process is driven by only one modality of condition. To further unleash the users' creativity, it is desirable for the model to be controllable by multiple modalities simultaneously, e.g., generating and editing faces by describing the age (text-driven) while drawing the face shape (mask-driven). In this work, we present Collaborative Diffusion, where pre-trained uni-modal diffusion models collaborate to achieve multi-modal face generation and editing without re-training. Our key insight is that diffusion models driven by different modalities are inherently complementary regarding the latent denoising steps, where bilateral connections can be established upon. Specifically, we propose dynamic diffuser, a meta-network that adaptively hallucinates multi-modal denoising steps by predicting the spatial-temporal influence functions for each pre-trained uni-modal model. Collaborative Diffusion not only collaborates generation capabilities from uni-modal diffusion models, but also integrates multiple uni-modal manipulations to perform multi-modal editing. Extensive qualitative and quantitative experiments demonstrate the superiority of our framework in both image quality and condition consistency.

Chenyang Si, Ziqi Huang, Yuming Jiang et al.

In this paper, we uncover the untapped potential of diffusion U-Net, which serves as a "free lunch" that substantially improves the generation quality on the fly. We initially investigate the key contributions of the U-Net architecture to the denoising process and identify that its main backbone primarily contributes to denoising, whereas its skip connections mainly introduce high-frequency features into the decoder module, causing the network to overlook the backbone semantics. Capitalizing on this discovery, we propose a simple yet effective method-termed "FreeU" - that enhances generation quality without additional training or finetuning. Our key insight is to strategically re-weight the contributions sourced from the U-Net's skip connections and backbone feature maps, to leverage the strengths of both components of the U-Net architecture. Promising results on image and video generation tasks demonstrate that our FreeU can be readily integrated to existing diffusion models, e.g., Stable Diffusion, DreamBooth, ModelScope, Rerender and ReVersion, to improve the generation quality with only a few lines of code. All you need is to adjust two scaling factors during inference. Project page: https://chenyangsi.top/FreeU/.

Ziqi Huang, Tianxing Wu, Yuming Jiang et al.

Diffusion models gain increasing popularity for their generative capabilities. Recently, there have been surging needs to generate customized images by inverting diffusion models from exemplar images, and existing inversion methods mainly focus on capturing object appearances (i.e., the "look"). However, how to invert object relations, another important pillar in the visual world, remains unexplored. In this work, we propose the Relation Inversion task, which aims to learn a specific relation (represented as "relation prompt") from exemplar images. Specifically, we learn a relation prompt with a frozen pre-trained text-to-image diffusion model. The learned relation prompt can then be applied to generate relation-specific images with new objects, backgrounds, and styles. To tackle the Relation Inversion task, we propose the ReVersion Framework. Specifically, we propose a novel "relation-steering contrastive learning" scheme to steer the relation prompt towards relation-dense regions, and disentangle it away from object appearances. We further devise "relation-focal importance sampling" to emphasize high-level interactions over low-level appearances (e.g., texture, color). To comprehensively evaluate this new task, we contribute the ReVersion Benchmark, which provides various exemplar images with diverse relations. Extensive experiments validate the superiority of our approach over existing methods across a wide range of visual relations. Our proposed task and method could be good inspirations for future research in various domains like generative inversion, few-shot learning, and visual relation detection.

Jianglin Fu, Shikai Li, Yuming Jiang et al.

Human generation has achieved significant progress. Nonetheless, existing methods still struggle to synthesize specific regions such as faces and hands. We argue that the main reason is rooted in the training data. A holistic human dataset inevitably has insufficient and low-resolution information on local parts. Therefore, we propose to use multi-source datasets with various resolution images to jointly learn a high-resolution human generative model. However, multi-source data inherently a) contains different parts that do not spatially align into a coherent human, and b) comes with different scales. To tackle these challenges, we propose an end-to-end framework, UnitedHuman, that empowers continuous GAN with the ability to effectively utilize multi-source data for high-resolution human generation. Specifically, 1) we design a Multi-Source Spatial Transformer that spatially aligns multi-source images to full-body space with a human parametric model. 2) Next, a continuous GAN is proposed with global-structural guidance and CutMix consistency. Patches from different datasets are then sampled and transformed to supervise the training of this scale-invariant generative model. Extensive experiments demonstrate that our model jointly learned from multi-source data achieves superior quality than those learned from a holistic dataset.

Yihang Chen, Tsai Hor Chan, Guosheng Yin et al.

Multiple instance learning (MIL) has been extensively applied to whole slide histopathology image (WSI) analysis. The existing aggregation strategy in MIL, which primarily relies on the first-order distance (e.g., mean difference) between instances, fails to accurately approximate the true feature distribution of each instance, leading to biased slide-level representations. Moreover, the scarcity of WSI observations easily leads to model overfitting, resulting in unstable testing performance and limited generalizability. To tackle these challenges, we propose a new Bayesian nonparametric framework for multiple instance learning, which adopts a cascade of Dirichlet processes (cDP) to incorporate the instance-to-bag characteristic of the WSIs. We perform feature aggregation based on the latent clusters formed by the Dirichlet process, which incorporates the covariances of the patch features and forms more representative clusters. We then perform bag-level prediction with another Dirichlet process model on the bags, which imposes a natural regularization on learning to prevent overfitting and enhance generalizability. Moreover, as a Bayesian nonparametric method, the cDP model can accurately generate posterior uncertainty, which allows for the detection of outlier samples and tumor localization. Extensive experiments on five WSI benchmarks validate the superior performance of our method, as well as its generalizability and ability to estimate uncertainties. Codes are available at https://github.com/HKU-MedAI/cDPMIL.

Haonan Qiu, Yuming Jiang, Hang Zhou et al.

Realistic generative face video synthesis has long been a pursuit in both computer vision and graphics community. However, existing face video generation methods tend to produce low-quality frames with drifted facial identities and unnatural movements. To tackle these challenges, we propose a principled framework named StyleFaceV, which produces high-fidelity identity-preserving face videos with vivid movements. Our core insight is to decompose appearance and pose information and recompose them in the latent space of StyleGAN3 to produce stable and dynamic results. Specifically, StyleGAN3 provides strong priors for high-fidelity facial image generation, but the latent space is intrinsically entangled. By carefully examining its latent properties, we propose our decomposition and recomposition designs which allow for the disentangled combination of facial appearance and movements. Moreover, a temporal-dependent model is built upon the decomposed latent features, and samples reasonable sequences of motions that are capable of generating realistic and temporally coherent face videos. Particularly, our pipeline is trained with a joint training strategy on both static images and high-quality video data, which is of higher data efficiency. Extensive experiments demonstrate that our framework achieves state-of-the-art face video generation results both qualitatively and quantitatively. Notably, StyleFaceV is capable of generating realistic $1024\times1024$ face videos even without high-resolution training videos.

Yuming Jiang, Nanxuan Zhao, Qing Liu et al.

Group portrait editing is highly desirable since users constantly want to add a person, delete a person, or manipulate existing persons. It is also challenging due to the intricate dynamics of human interactions and the diverse gestures. In this work, we present GroupDiff, a pioneering effort to tackle group photo editing with three dedicated contributions: 1) Data Engine: Since there is no labeled data for group photo editing, we create a data engine to generate paired data for training. The training data engine covers the diverse needs of group portrait editing. 2) Appearance Preservation: To keep the appearance consistent after editing, we inject the images of persons from the group photo into the attention modules and employ skeletons to provide intra-person guidance. 3) Control Flexibility: Bounding boxes indicating the locations of each person are used to reweight the attention matrix so that the features of each person can be injected into the correct places. This inter-person guidance provides flexible manners for manipulation. Extensive experiments demonstrate that GroupDiff exhibits state-of-the-art performance compared to existing methods. GroupDiff offers controllability for editing and maintains the fidelity of the original photos.

Chulong Zhang, Yuming Jiang, Na Li et al.

Pathologists need to combine information from differently stained pathology slices for accurate diagnosis. Deformable image registration is a necessary technique for fusing multi-modal pathology slices. This paper proposes a hybrid deep feature-based deformable image registration framework for stained pathology samples. We first extract dense feature points via the detector-based and detector-free deep learning feature networks and perform points matching. Then, to further reduce false matches, an outlier detection method combining the isolation forest statistical model and the local affine correction model is proposed. Finally, the interpolation method generates the deformable vector field for pathology image registration based on the above matching points. We evaluate our method on the dataset of the Non-rigid Histology Image Registration (ANHIR) challenge, which is co-organized with the IEEE ISBI 2019 conference. Our technique outperforms the traditional approaches by 17% with the Average-Average registration target error (rTRE) reaching 0.0034. The proposed method achieved state-of-the-art performance and ranked 1st in evaluating the test dataset. The proposed hybrid deep feature-based registration method can potentially become a reliable method for pathology image registration.

Ziqi Huang, Fan Zhang, Xiaojie Xu et al.

Video generation has witnessed significant advancements, yet evaluating these models remains a challenge. A comprehensive evaluation benchmark for video generation is indispensable for two reasons: 1) Existing metrics do not fully align with human perceptions; 2) An ideal evaluation system should provide insights to inform future developments of video generation. To this end, we present VBench, a comprehensive benchmark suite that dissects "video generation quality" into specific, hierarchical, and disentangled dimensions, each with tailored prompts and evaluation methods. VBench has several appealing properties: 1) Comprehensive Dimensions: VBench comprises 16 dimensions in video generation (e.g., subject identity inconsistency, motion smoothness, temporal flickering, and spatial relationship, etc). The evaluation metrics with fine-grained levels reveal individual models' strengths and weaknesses. 2) Human Alignment: We also provide a dataset of human preference annotations to validate our benchmarks' alignment with human perception, for each evaluation dimension respectively. 3) Valuable Insights: We look into current models' ability across various evaluation dimensions, and various content types. We also investigate the gaps between video and image generation models. 4) Versatile Benchmarking: VBench++ supports evaluating text-to-video and image-to-video. We introduce a high-quality Image Suite with an adaptive aspect ratio to enable fair evaluations across different image-to-video generation settings. Beyond assessing technical quality, VBench++ evaluates the trustworthiness of video generative models, providing a more holistic view of model performance. 5) Full Open-Sourcing: We fully open-source VBench++ and continually add new video generation models to our leaderboard to drive forward the field of video generation.

Lin Liu, Xinxin Fan, Haoyang Liu et al.

Rigid pre-registration involving local-global matching or other large deformation scenarios is crucial. Current popular methods rely on unsupervised learning based on grayscale similarity, but under circumstances where different poses lead to varying tissue structures, or where image quality is poor, these methods tend to exhibit instability and inaccuracies. In this study, we propose a novel method for medical image registration based on arbitrary voxel point of interest matching, called query point quizzer (QUIZ). QUIZ focuses on the correspondence between local-global matching points, specifically employing CNN for feature extraction and utilizing the Transformer architecture for global point matching queries, followed by applying average displacement for local image rigid transformation. We have validated this approach on a large deformation dataset of cervical cancer patients, with results indicating substantially smaller deviations compared to state-of-the-art methods. Remarkably, even for cross-modality subjects, it achieves results surpassing the current state-of-the-art.

Sicong Leng, Jing Wang, Jiaxi Li et al.

Large multimodal reasoning models have achieved rapid progress, but their advancement is constrained by two major limitations: the absence of open, large-scale, high-quality long chain-of-thought (CoT) data, and the instability of reinforcement learning (RL) algorithms in post-training. Group Relative Policy Optimization (GRPO), the standard framework for RL fine-tuning, is prone to gradient vanishing when reward variance is low, which weakens optimization signals and impairs convergence. This work makes three contributions: (1) We propose Variance-Aware Sampling (VAS), a data selection strategy guided by Variance Promotion Score (VPS) that combines outcome variance and trajectory diversity to promote reward variance and stabilize policy optimization. (2) We release large-scale, carefully curated resources containing ~1.6M long CoT cold-start data and ~15k RL QA pairs, designed to ensure quality, difficulty, and diversity, along with a fully reproducible end-to-end training codebase. (3) We open-source a family of multimodal reasoning models in multiple scales, establishing standardized baselines for the community. Experiments across mathematical reasoning benchmarks demonstrate the effectiveness of both the curated data and the proposed VAS. Comprehensive ablation studies and analyses provide further insight into the contributions of each component. In addition, we theoretically establish that reward variance lower-bounds the expected policy gradient magnitude, with VAS serving as a practical mechanism to realize this guarantee. Our code, data, and checkpoints are available at https://github.com/LengSicong/MMR1.

Rui Cao, Yuming Jiang, Michael Schlichtkrull et al.

Multimodal Large Language Models (MLLMs) can enhance trustworthiness by aligning with human preferences. As human preference labeling is laborious, recent works employ evaluation models for assessing MLLMs' responses, using the model-based assessments to automate preference dataset construction. This approach, however, faces challenges with MLLMs' lengthy and compositional responses, which often require diverse reasoning skills that a single evaluation model may not fully possess. Additionally, most existing methods rely on closed-source models as evaluators. To address limitations, we propose DecompGen, a decomposable framework that uses an ensemble of open-sourced expert models. DecompGen breaks down each response into atomic verification tasks, assigning each task to an appropriate expert model to generate fine-grained assessments. The DecompGen feedback is used to automatically construct our preference dataset, DGPref. MLLMs aligned with DGPref via preference learning show improvements in trustworthiness, demonstrating the effectiveness of DecompGen.

Boqiang Zhang, Kehan Li, Zesen Cheng et al. · pku

In this paper, we propose VideoLLaMA3, a more advanced multimodal foundation model for image and video understanding. The core design philosophy of VideoLLaMA3 is vision-centric. The meaning of "vision-centric" is two-fold: the vision-centric training paradigm and vision-centric framework design. The key insight of our vision-centric training paradigm is that high-quality image-text data is crucial for both image and video understanding. Instead of preparing massive video-text datasets, we focus on constructing large-scale and high-quality image-text datasets. VideoLLaMA3 has four training stages: 1) Vision Encoder Adaptation, which enables vision encoder to accept images of variable resolutions as input; 2) Vision-Language Alignment, which jointly tunes the vision encoder, projector, and LLM with large-scale image-text data covering multiple types (including scene images, documents, charts) as well as text-only data. 3) Multi-task Fine-tuning, which incorporates image-text SFT data for downstream tasks and video-text data to establish a foundation for video understanding. 4) Video-centric Fine-tuning, which further improves the model's capability in video understanding. As for the framework design, to better capture fine-grained details in images, the pretrained vision encoder is adapted to encode images of varying sizes into vision tokens with corresponding numbers, rather than a fixed number of tokens. For video inputs, we reduce the number of vision tokens according to their similarity so that the representation of videos will be more precise and compact. Benefit from vision-centric designs, VideoLLaMA3 achieves compelling performances in both image and video understanding benchmarks.

Chao Li, Dasha Hu, Chengyang Li et al.

Unsupervised Domain Adaptation transfers knowledge from a labeled source domain to an unlabeled target domain. Directly deploying Vision-Language Models (VLMs) with prompt tuning in downstream UDA tasks faces the signifi cant challenge of mitigating domain discrepancies. Existing prompt-tuning strategies primarily align marginal distribu tion, but neglect conditional distribution discrepancies, lead ing to critical issues such as class prototype misalignment and degraded semantic discriminability. To address these lim itations, the work proposes C-DGPA: Class-Centric Dual Alignment Generative Prompt Adaptation. C-DGPA syner gistically optimizes marginal distribution alignment and con ditional distribution alignment through a novel dual-branch architecture. The marginal distribution alignment branch em ploys a dynamic adversarial training framework to bridge marginal distribution discrepancies. Simultaneously, the con ditional distribution alignment branch introduces a Class Mapping Mechanism (CMM) to align conditional distribu tion discrepancies by standardizing semantic prompt under standing and preventing source domain over-reliance. This dual alignment strategy effectively integrates domain knowl edge into prompt learning via synergistic optimization, ensur ing domain-invariant and semantically discriminative repre sentations. Extensive experiments on OfficeHome, Office31, and VisDA-2017 validate the superiority of C-DGPA. It achieves new state-of-the-art results on all benchmarks.

Tianxing Wu, Chenyang Si, Yuming Jiang et al.

Though diffusion-based video generation has witnessed rapid progress, the inference results of existing models still exhibit unsatisfactory temporal consistency and unnatural dynamics. In this paper, we delve deep into the noise initialization of video diffusion models, and discover an implicit training-inference gap that attributes to the unsatisfactory inference quality.Our key findings are: 1) the spatial-temporal frequency distribution of the initial noise at inference is intrinsically different from that for training, and 2) the denoising process is significantly influenced by the low-frequency components of the initial noise. Motivated by these observations, we propose a concise yet effective inference sampling strategy, FreeInit, which significantly improves temporal consistency of videos generated by diffusion models. Through iteratively refining the spatial-temporal low-frequency components of the initial latent during inference, FreeInit is able to compensate the initialization gap between training and inference, thus effectively improving the subject appearance and temporal consistency of generation results. Extensive experiments demonstrate that FreeInit consistently enhances the generation quality of various text-to-video diffusion models without additional training or fine-tuning.

Long Li, Weiwen Xu, Jiayan Guo et al. · pku

Effective research ideation is a critical step for scientific research. However, the exponential increase in scientific literature makes it challenging for researchers to stay current with recent advances and identify meaningful research directions. Recent developments in large language models~(LLMs) suggest a promising avenue for automating the generation of novel research ideas. However, existing methods for idea generation either trivially prompt LLMs or directly expose LLMs to extensive literature without indicating useful information. Inspired by the research process of human researchers, we propose a Chain-of-Ideas~(CoI) agent, an LLM-based agent that organizes relevant literature in a chain structure to effectively mirror the progressive development in a research domain. This organization facilitates LLMs to capture the current advancements in research, thereby enhancing their ideation capabilities. Furthermore, we propose Idea Arena, an evaluation protocol that can comprehensively evaluate idea generation methods from different perspectives, aligning closely with the preferences of human researchers. Experimental results indicate that the CoI agent consistently outperforms other methods and shows comparable quality as humans in research idea generation. Moreover, our CoI agent is budget-friendly, with a minimum cost of \$0.50 to generate a candidate idea and its corresponding experimental design.

Jun Cen, Chaohui Yu, Hangjie Yuan et al. · pku

We present WorldVLA, an autoregressive action world model that unifies action and image understanding and generation. Our WorldVLA intergrates Vision-Language-Action (VLA) model and world model in one single framework. The world model predicts future images by leveraging both action and image understanding, with the purpose of learning the underlying physics of the environment to improve action generation. Meanwhile, the action model generates the subsequent actions based on image observations, aiding in visual understanding and in turn helps visual generation of the world model. We demonstrate that WorldVLA outperforms standalone action and world models, highlighting the mutual enhancement between the world model and the action model. In addition, we find that the performance of the action model deteriorates when generating sequences of actions in an autoregressive manner. This phenomenon can be attributed to the model's limited generalization capability for action prediction, leading to the propagation of errors from earlier actions to subsequent ones. To address this issue, we propose an attention mask strategy that selectively masks prior actions during the generation of the current action, which shows significant performance improvement in the action chunk generation task.

Weichen Fan, Chenyang Si, Junhao Song et al.

We present Vchitect-2.0, a parallel transformer architecture designed to scale up video diffusion models for large-scale text-to-video generation. The overall Vchitect-2.0 system has several key designs. (1) By introducing a novel Multimodal Diffusion Block, our approach achieves consistent alignment between text descriptions and generated video frames, while maintaining temporal coherence across sequences. (2) To overcome memory and computational bottlenecks, we propose a Memory-efficient Training framework that incorporates hybrid parallelism and other memory reduction techniques, enabling efficient training of long video sequences on distributed systems. (3) Additionally, our enhanced data processing pipeline ensures the creation of Vchitect T2V DataVerse, a high-quality million-scale training dataset through rigorous annotation and aesthetic evaluation. Extensive benchmarking demonstrates that Vchitect-2.0 outperforms existing methods in video quality, training efficiency, and scalability, serving as a suitable base for high-fidelity video generation.

Qing Lyu, Jeremy Hudson, Mohammad Kawas et al.

Individualized Alzheimer's disease (AD) progression prediction requires models that use irregular visits, account for censoring, avoid diagnostic leakage, and provide calibrated horizon risks. We propose PROgression-aware MultI-horizon Survival Estimation for Alzheimer's Disease (PROMISE-AD), a leakage-safe survival framework for predicting conversion from cognitively normal (CN) to mild cognitive impairment (MCI) and from MCI to AD dementia using ADNI/TADPOLE tabular histories. PROMISE-AD converts pre-index visits into tokens with standardized measurements, missingness masks, longitudinal changes, time-normalized slopes, visit timing, and non-diagnostic categorical attributes. A temporal Transformer fuses global, attention-pooled, and latest-visit representations to estimate a progression score and latent discrete-time mixture hazards. Training combines survival likelihood, horizon-specific focal risk loss, progression ranking, hazard smoothness, and mixture-balance regularization, followed by validation-set isotonic calibration for 1-, 2-, 3-, and 5-year risks. In held-out testing across three seeds, PROMISE-AD achieved an integrated Brier score (IBS) of 0.085 $\pm$ 0.012, C-index of 0.808 $\pm$ 0.015, and mean time-dependent AUC of 0.840 $\pm$ 0.081 for CN-to-MCI conversion, yielding the lowest IBS among compared methods. For MCI-to-AD conversion, PROMISE-AD achieved the highest C-index (0.894 $\pm$ 0.018) and near-ceiling 5-year discrimination (AUROC 0.997 $\pm$ 0.003; AUPRC 0.999 $\pm$ 0.001), although some baselines had lower IBS. Ablations and interpretability supported longitudinal change features, fused temporal representations, mixture hazards, cognitive and functional measures, APOE4 status, and recent conversion-proximal visits. These findings suggest that progression-aware survival modeling can provide interpretable multi-horizon AD conversion risk estimates.

Weiqin Zhao, Ziyu Guo, Yinshuang Fan et al.

Due to the large size and lack of fine-grained annotation, Whole Slide Images (WSIs) analysis is commonly approached as a Multiple Instance Learning (MIL) problem. However, previous studies only learn from training data, posing a stark contrast to how human clinicians teach each other and reason about histopathologic entities and factors. Here we present a novel knowledge concept-based MIL framework, named ConcepPath to fill this gap. Specifically, ConcepPath utilizes GPT-4 to induce reliable diseasespecific human expert concepts from medical literature, and incorporate them with a group of purely learnable concepts to extract complementary knowledge from training data. In ConcepPath, WSIs are aligned to these linguistic knowledge concepts by utilizing pathology vision-language model as the basic building component. In the application of lung cancer subtyping, breast cancer HER2 scoring, and gastric cancer immunotherapy-sensitive subtyping task, ConcepPath significantly outperformed previous SOTA methods which lack the guidance of human expert knowledge.

Yuming Jiang, Siteng Huang, Shengke Xue et al. · pku

This paper presents RynnVLA-001, a vision-language-action(VLA) model built upon large-scale video generative pretraining from human demonstrations. We propose a novel two-stage pretraining methodology. The first stage, Ego-Centric Video Generative Pretraining, trains an Image-to-Video model on 12M ego-centric manipulation videos to predict future frames conditioned on an initial frame and a language instruction. The second stage, Human-Centric Trajectory-Aware Modeling, extends this by jointly predicting future keypoint trajectories, thereby effectively bridging visual frame prediction with action prediction. Furthermore, to enhance action representation, we propose ActionVAE, a variational autoencoder that compresses sequences of actions into compact latent embeddings, reducing the complexity of the VLA output space. When finetuned on the same downstream robotics datasets, RynnVLA-001 achieves superior performance over state-of-the-art baselines, demonstrating that the proposed pretraining strategy provides a more effective initialization for VLA models.

Chunhan Li, Qifeng Wu, Jia-Hui Pan et al.

Text-to-image generation has advanced rapidly, but existing models still struggle with faithfully composing multiple objects and preserving their attributes in complex scenes. We propose coDrawAgents, an interactive multi-agent dialogue framework with four specialized agents: Interpreter, Planner, Checker, and Painter that collaborate to improve compositional generation. The Interpreter adaptively decides between a direct text-to-image pathway and a layout-aware multi-agent process. In the layout-aware mode, it parses the prompt into attribute-rich object descriptors, ranks them by semantic salience, and groups objects with the same semantic priority level for joint generation. Guided by the Interpreter, the Planner adopts a divide-and-conquer strategy, incrementally proposing layouts for objects with the same semantic priority level while grounding decisions in the evolving visual context of the canvas. The Checker introduces an explicit error-correction mechanism by validating spatial consistency and attribute alignment, and refining layouts before they are rendered. Finally, the Painter synthesizes the image step by step, incorporating newly planned objects into the canvas to provide richer context for subsequent iterations. Together, these agents address three key challenges: reducing layout complexity, grounding planning in visual context, and enabling explicit error correction. Extensive experiments on benchmarks GenEval and DPG-Bench demonstrate that coDrawAgents substantially improves text-image alignment, spatial accuracy, and attribute binding compared to existing methods.

Liang Qiu, Wenhao Chi, Xiaohan Xing et al.

Precision therapy for liver cancer necessitates accurately delineating liver sub-regions to protect healthy tissue while targeting tumors, which is essential for reducing recurrence and improving survival rates. However, the segmentation of hepatic segments, known as Couinaud segmentation, is challenging due to indistinct sub-region boundaries and the need for extensive annotated datasets. This study introduces LiverFormer, a novel Couinaud segmentation model that effectively integrates global context with low-level local features based on a 3D hybrid CNN-Transformer architecture. Additionally, a registration-based data augmentation strategy is equipped to enhance the segmentation performance with limited labeled data. Evaluated on CT images from 123 patients, LiverFormer demonstrated high accuracy and strong concordance with expert annotations across various metrics, allowing for enhanced treatment planning for surgery and radiation therapy. It has great potential to reduces complications and minimizes potential damages to surrounding tissue, leading to improved outcomes for patients undergoing complex liver cancer treatments.

Befekadu G. Gebraselase, Bjarne E. Helvik, Yuming Jiang

Blockchain is a technology that provides a distributed ledger that stores previous records while maintaining consistency and security. Bitcoin is the first and largest decentralized electronic cryptographic system that uses blockchain technology. It faces a challenge in making all the nodes synchronize and have the same overall view with the cost of scalability and performance. In addition, with miners' financial interest playing a significant role in choosing transactions from the backlog, small fee or small fee per byte value transactions will exhibit more delays. To study the issues related to the system's performance, we developed an $M(t)/M^N/1$ model. The backlog's arrival follows an inhomogeneous Poison process to the system that has infinite buffer capacity, and the service time is distributed exponentially, which removes $N$ transactions at time. Besides validating the model with measurement data, we have used the model to study the reward distribution when miners take transaction selection strategies like fee per byte, fee-based, and FIFO. The analysis shows that smaller fee transactions exhibit higher waiting times, even with increasing the block size. Moreover, the miner transaction selection strategy impacts the final gain.

Yuming Jiang, Ziqi Huang, Xingang Pan et al.

Facial editing is an important task in vision and graphics with numerous applications. However, existing works are incapable to deliver a continuous and fine-grained editing mode (e.g., editing a slightly smiling face to a big laughing one) with natural interactions with users. In this work, we propose Talk-to-Edit, an interactive facial editing framework that performs fine-grained attribute manipulation through dialog between the user and the system. Our key insight is to model a continual "semantic field" in the GAN latent space. 1) Unlike previous works that regard the editing as traversing straight lines in the latent space, here the fine-grained editing is formulated as finding a curving trajectory that respects fine-grained attribute landscape on the semantic field. 2) The curvature at each step is location-specific and determined by the input image as well as the users' language requests. 3) To engage the users in a meaningful dialog, our system generates language feedback by considering both the user request and the current state of the semantic field. We also contribute CelebA-Dialog, a visual-language facial editing dataset to facilitate large-scale study. Specifically, each image has manually annotated fine-grained attribute annotations as well as template-based textual descriptions in natural language. Extensive quantitative and qualitative experiments demonstrate the superiority of our framework in terms of 1) the smoothness of fine-grained editing, 2) the identity/attribute preservation, and 3) the visual photorealism and dialog fluency. Notably, user study validates that our overall system is consistently favored by around 80% of the participants. Our project page is https://www.mmlab-ntu.com/project/talkedit/.

Befekadu G. Gebraselase, Bjarne E. Helvik, Yuming Jiang

Bitcoin has become the leading cryptocurrency system, but the limit on its transaction processing capacity has resulted in increased transaction fees and delayed transaction confirmation. As such, it is pertinent to understand and probably predict how transactions are handled by Bitcoin such that a user may adapt the transaction requests and a miner may adjust the block generation strategy and/or the mining pool to join. To this aim, the present paper introduces results from an analysis of transaction handling in Bitcoin. Specifically, the analysis consists of two-part. The first part is an exploratory data analysis revealing key characteristics in Bitcoin transaction handling. The second part is a predictability analysis intended to provide insights on transaction handling such as (i) transaction confirmation time, (ii) block attributes, and (iii) who has created the block. The result shows that some models do reasonably well for (ii), but surprisingly not for (i) or (iii).

Yuming Jiang, Kelvin C. K. Chan, Xintao Wang et al.

Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g. scale and rotation) and the resolution gap (e.g. HR and LR). To tackle these challenges, we propose C2-Matching in this work, which produces explicit robust matching crossing transformation and resolution. 1) For the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) For the resolution gap, we adopt a teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue. In addition, to faithfully evaluate the performance of Ref-SR under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. Extensive experiments demonstrate that our proposed C2-Matching significantly outperforms state of the arts by over 1dB on the standard CUFED5 benchmark. Notably, it also shows great generalizability on WR-SR dataset as well as robustness across large scale and rotation transformations.

Befekadu G. Gebraselase, Bjarne E. Helvik, Yuming Jiang

Blockchain has been considered as an important technique to enable secure management of virtual network functions and network slices. To understand such capabilities of a blockchain, e.g. transaction confirmation time, demands a thorough study on the transaction characteristics of the blockchain. This paper presents a comprehensive study on the transaction characteristics of Bitcoin -- the first blockchain application, focusing on the underlying fundamental processes. A set of results and findings are obtained, which provide new insight into understanding the transaction and traffic characteristics of Bitcoin. As a highlight, the validity of several hypotheses/assumptions used in the literature is examined with measurement for the first time.