Wei Ji

Ao Zhang, Yuan Yao, Qianyu Chen et al.

Scene graph generation (SGG) is designed to extract (subject, predicate, object) triplets in images. Recent works have made a steady progress on SGG, and provide useful tools for high-level vision and language understanding. However, due to the data distribution problems including long-tail distribution and semantic ambiguity, the predictions of current SGG models tend to collapse to several frequent but uninformative predicates (e.g., on, at), which limits practical application of these models in downstream tasks. To deal with the problems above, we propose a novel Internal and External Data Transfer (IETrans) method, which can be applied in a plug-and-play fashion and expanded to large SGG with 1,807 predicate classes. Our IETrans tries to relieve the data distribution problem by automatically creating an enhanced dataset that provides more sufficient and coherent annotations for all predicates. By training on the enhanced dataset, a Neural Motif model doubles the macro performance while maintaining competitive micro performance. The code and data are publicly available at https://github.com/waxnkw/IETrans-SGG.pytorch.

Junde Wu, Wei Ji, Yuanpei Liu et al.

The Segment Anything Model (SAM) has recently gained popularity in the field of image segmentation due to its impressive capabilities in various segmentation tasks and its prompt-based interface. However, recent studies and individual experiments have shown that SAM underperforms in medical image segmentation, since the lack of the medical specific knowledge. This raises the question of how to enhance SAM's segmentation capability for medical images. In this paper, instead of fine-tuning the SAM model, we propose the Medical SAM Adapter (Med-SA), which incorporates domain-specific medical knowledge into the segmentation model using a light yet effective adaptation technique. In Med-SA, we propose Space-Depth Transpose (SD-Trans) to adapt 2D SAM to 3D medical images and Hyper-Prompting Adapter (HyP-Adpt) to achieve prompt-conditioned adaptation. We conduct comprehensive evaluation experiments on 17 medical image segmentation tasks across various image modalities. Med-SA outperforms several state-of-the-art (SOTA) medical image segmentation methods, while updating only 2\% of the parameters. Our code is released at https://github.com/KidsWithTokens/Medical-SAM-Adapter.

Yuan Yao, Qianyu Chen, Ao Zhang et al.

Vision-language pre-training (VLP) has shown impressive performance on a wide range of cross-modal tasks, where VLP models without reliance on object detectors are becoming the mainstream due to their superior computation efficiency and competitive performance. However, the removal of object detectors also deprives the capability of VLP models in explicit object modeling, which is essential to various position-sensitive vision-language (VL) tasks, such as referring expression comprehension and visual commonsense reasoning. To address the challenge, we introduce PEVL that enhances the pre-training and prompt tuning of VLP models with explicit object position modeling. Specifically, PEVL reformulates discretized object positions and language in a unified language modeling framework, which facilitates explicit VL alignment during pre-training, and also enables flexible prompt tuning for various downstream tasks. We show that PEVL enables state-of-the-art performance of detector-free VLP models on position-sensitive tasks such as referring expression comprehension and phrase grounding, and also improves the performance on position-insensitive tasks with grounded inputs. We make the data and code for this paper publicly available at https://github.com/thunlp/PEVL.

Juncheng Li, Kaihang Pan, Zhiqi Ge et al.

Recent advancements in Multimodal Large Language Models (MLLMs) have been utilizing Visual Prompt Generators (VPGs) to convert visual features into tokens that LLMs can recognize. This is achieved by training the VPGs on millions of image-caption pairs, where the VPG-generated tokens of images are fed into a frozen LLM to generate the corresponding captions. However, this image-captioning based training objective inherently biases the VPG to concentrate solely on the primary visual contents sufficient for caption generation, often neglecting other visual details. This shortcoming results in MLLMs' underperformance in comprehending demonstrative instructions consisting of multiple, interleaved, and multimodal instructions that demonstrate the required context to complete a task. To address this issue, we introduce a generic and lightweight Visual Prompt Generator Complete module (VPG-C), which can infer and complete the missing details essential for comprehending demonstrative instructions. Further, we propose a synthetic discriminative training strategy to fine-tune VPG-C, eliminating the need for supervised demonstrative instructions. As for evaluation, we build DEMON, a comprehensive benchmark for demonstrative instruction understanding. Synthetically trained with the proposed strategy, VPG-C achieves significantly stronger zero-shot performance across all tasks of DEMON. Further evaluation on the MME and OwlEval benchmarks also demonstrate the superiority of VPG-C. Our benchmark, code, and pre-trained models are available at https://github.com/DCDmllm/Cheetah.

Yali Du, Yinwei Wei, Wei Ji et al.

The booming development and huge market of micro-videos bring new e-commerce channels for merchants. Currently, more micro-video publishers prefer to embed relevant ads into their micro-videos, which not only provides them with business income but helps the audiences to discover their interesting products. However, due to the micro-video recording by unprofessional equipment, involving various topics and including multiple modalities, it is challenging to locate the products related to micro-videos efficiently, appropriately, and accurately. We formulate the microvideo-product retrieval task, which is the first attempt to explore the retrieval between the multi-modal and multi-modal instances. A novel approach named Multi-Queue Momentum Contrast (MQMC) network is proposed for bidirectional retrieval, consisting of the uni-modal feature and multi-modal instance representation learning. Moreover, a discriminative selection strategy with a multi-queue is used to distinguish the importance of different negatives based on their categories. We collect two large-scale microvideo-product datasets (MVS and MVS-large) for evaluation and manually construct the hierarchical category ontology, which covers sundry products in daily life. Extensive experiments show that MQMC outperforms the state-of-the-art baselines. Our replication package (including code, dataset, etc.) is publicly available at https://github.com/duyali2000/MQMC.

Xiang Fang, Wanlong Fang, Changshuo Wang et al.

Video-Language Models (VLMs) have demonstrated impressive multi-modal reasoning capabilities across diverse computer vision applications. However, these VLMs are task-specific and assume that both video and language inputs are complete. However, real-world VLM applications might face challenges due to deactivated sensors (e.g., cameras are unavailable due to data privacy), yielding modality-incomplete data and leading to inconsistency between training and testing data. While straightforward incomplete input can boast training generalization-ability and lead to training failure, its potential risks to VLMs regarding safety and trustworthiness have been largely neglected. To this end, we make the first attempt to propose a unified incomplete video-language model to process the incomplete multi-modal inputs. Extensive experimental results show that our method can serve as a plug-and-play module for previous works to improve their performance in various multi-modal tasks.

Shengqiong Wu, Hao Fei, Leigang Qu et al.

While recently Multimodal Large Language Models (MM-LLMs) have made exciting strides, they mostly fall prey to the limitation of only input-side multimodal understanding, without the ability to produce content in multiple modalities. As we humans always perceive the world and communicate with people through various modalities, developing any-to-any MM-LLMs capable of accepting and delivering content in any modality becomes essential to human-level AI. To fill the gap, we present an end-to-end general-purpose any-to-any MM-LLM system, NExT-GPT. We connect an LLM with multimodal adaptors and different diffusion decoders, enabling NExT-GPT to perceive inputs and generate outputs in arbitrary combinations of text, images, videos, and audio. By leveraging the existing well-trained highly-performing encoders and decoders, NExT-GPT is tuned with only a small amount of parameter (1%) of certain projection layers, which not only benefits low-cost training and also facilitates convenient expansion to more potential modalities. Moreover, we introduce a modality-switching instruction tuning (MosIT) and manually curate a high-quality dataset for MosIT, based on which NExT-GPT is empowered with complex cross-modal semantic understanding and content generation. Overall, our research showcases the promising possibility of building an AI agent capable of modeling universal modalities, paving the way for more human-like AI research in the community. Project page: https://next-gpt.github.io/

Ao Zhang, Yuan Yao, Wei Ji et al.

The development of large language models (LLMs) has greatly advanced the field of multimodal understanding, leading to the emergence of large multimodal models (LMMs). In order to enhance the level of visual comprehension, recent studies have equipped LMMs with region-level understanding capabilities by representing object bounding box coordinates as a series of text sequences (pix2seq). In this paper, we introduce a novel paradigm for object location modeling called pix2emb method, where we ask the LMM to output the location embeddings and then decode them with different decoders. This paradigm allows us to use different location formats (such as bounding boxes and masks) in multimodal conversations. Leveraging the proposed pix2emb method, we train an LMM named NExT-Chat and demonstrate its capability of handling multiple tasks like visual grounding, region captioning, and grounded reasoning. Comprehensive experiments show the effectiveness of our NExT-Chat on various tasks, e.g., NExT-Chat (87.7) vs. Shikra (86.9) on POPE-Random, NExT-Chat (68.9) vs. LISA (67.9) on referring expression segmentation task, and NExT-Chat (79.6) vs. Kosmos-2 (62.3) on region caption task. The code and model are released at https://github.com/NExT-ChatV/NExT-Chat.

Junde Wu, Wei Ji, Huazhu Fu et al.

The Diffusion Probabilistic Model (DPM) has recently gained popularity in the field of computer vision, thanks to its image generation applications, such as Imagen, Latent Diffusion Models, and Stable Diffusion, which have demonstrated impressive capabilities and sparked much discussion within the community. Recent investigations have further unveiled the utility of DPM in the domain of medical image analysis, as underscored by the commendable performance exhibited by the medical image segmentation model across various tasks. Although these models were originally underpinned by a UNet architecture, there exists a potential avenue for enhancing their performance through the integration of vision transformer mechanisms. However, we discovered that simply combining these two models resulted in subpar performance. To effectively integrate these two cutting-edge techniques for the Medical image segmentation, we propose a novel Transformer-based Diffusion framework, called MedSegDiff-V2. We verify its effectiveness on 20 medical image segmentation tasks with different image modalities. Through comprehensive evaluation, our approach demonstrates superiority over prior state-of-the-art (SOTA) methodologies. Code is released at https://github.com/KidsWithTokens/MedSegDiff

Xiang Fang, Wanlong Fang, Wei Ji

Large Vision-Language Models have achieved unprecedented success in zero-shot recognition by aligning visual features with broad semantic concepts. However, this semantic abstraction creates a critical vulnerability in open-world deployment: the ``Hubris of Semantics'', where models force-fit unknown anomalies into known categories with high confidence due to the lack of explicit negative knowledge. To address this \textit{Open-World Trustworthiness Paradox}, we propose \textbf{Immuno-VLM}, a bio-inspired framework that adapts the biological principle of \textbf{Immunological Negative Selection} to high-dimensional latent spaces. Departing from traditional Open-Set Recognition methods that rely on passive density estimation or inefficient pixel-space outlier generation, Immuno-VLM leverages the generative reasoning of Large Language Models to actively hallucinate ``Semantic Antibodies'', textual descriptions of near-distribution outliers (e.g., look-alikes, contextual anomalies) that effectively bound the decision space of known classes.Extensive experiments on ImageNet-1K and four challenging OOD benchmarks reveal that Immuno-VLM establishes a new state-of-the-art.

Shihang Zhang, Mingjin Kuai, Ye Wei et al.

Video Moment Retrieval (VMR) task requires accurately localizing temporal boundaries aligned with natural language queries, but many models suffer from a misalignment between continuous surrogate losses and non-differentiable metrics, leading to optimization stagnation during the late stages of training and trapping boundary predictions in suboptimal solutions. Although Reinforcement Learning (RL) post-training successfully optimizes localization results for large models, applying it directly to lightweight networks easily disrupts the fragile feature representations established during the supervised phase. To overcome this optimization bottleneck, we propose Gradient-Isolated Reinforcement Learning for DETR (GIRL-DETR), introducing RL post-training into a lightweight temporal localization framework for the first time. The input video and text features first establish early alignment through Cross-Modal Interaction (CMI) before entering the transformer encoder. Subsequently, a Text-Guided Gating (TGG) mechanism dynamically injects semantic priors into the queries before the transformer decoder generates candidate proposals, providing high signal-to-noise ratio inputs for temporal prediction. After the supervised training reaches convergence, the backbone network is frozen to protect the feature manifold, while the detection head directly optimizes the non-differentiable evaluation metric tIoU to enhance localization accuracy through a Three-stage Progressive Reinforcement Learning (TPRL) strategy. This approach achieves an orthogonal decoupling of state representation and metric optimization. Experiments on Charades-STA, QVHighlights, and TACoS demonstrate that GIRL-DETR effectively resolves surrogate loss degradation and achieves substantial accuracy improvements with minimal parameter updates, providing a robust new pathway for RL applications in lightweight VMR models.

Zhedong Zheng, Jiayin Zhu, Wei Ji et al.

This research aims to study a self-supervised 3D clothing reconstruction method, which recovers the geometry shape and texture of human clothing from a single image. Compared with existing methods, we observe that three primary challenges remain: (1) 3D ground-truth meshes of clothing are usually inaccessible due to annotation difficulties and time costs; (2) Conventional template-based methods are limited to modeling non-rigid objects, e.g., handbags and dresses, which are common in fashion images; (3) The inherent ambiguity compromises the model training, such as the dilemma between a large shape with a remote camera or a small shape with a close camera. In an attempt to address the above limitations, we propose a causality-aware self-supervised learning method to adaptively reconstruct 3D non-rigid objects from 2D images without 3D annotations. In particular, to solve the inherent ambiguity among four implicit variables, i.e., camera position, shape, texture, and illumination, we introduce an explainable structural causal map (SCM) to build our model. The proposed model structure follows the spirit of the causal map, which explicitly considers the prior template in the camera estimation and shape prediction. When optimization, the causality intervention tool, i.e., two expectation-maximization loops, is deeply embedded in our algorithm to (1) disentangle four encoders and (2) facilitate the prior template. Extensive experiments on two 2D fashion benchmarks (ATR and Market-HQ) show that the proposed method could yield high-fidelity 3D reconstruction. Furthermore, we also verify the scalability of the proposed method on a fine-grained bird dataset, i.e., CUB. The code is available at https://github.com/layumi/ 3D-Magic-Mirror .

Xu Pan, Qiyuan Ma, Mingyue Dong et al.

Reliable correspondence estimation is a fundamental problem in image processing, underpinning applications such as Structure from Motion, visual localization, and image registration. Existing learning-based methods have significantly improved local feature representations, yet most still operate at the pixel or patch level and lack explicit modeling of regions that are jointly visible across views. We propose SAMatcher, a feature matching framework that formulates correspondence estimation through co-visibility modeling. Instead of directly matching local features, SAMatcher first predicts co-visible region masks and bounding boxes as structured priors for correspondence estimation. Built upon the Segment Anything Model (SAM), it introduces a symmetric cross-view interaction mechanism that enables bidirectional feature exchange and cross-view semantic alignment. We further develop a unified supervision scheme that jointly optimizes mask prediction and box localization through mask learning, box regression, and mask-box consistency constraints. Extensive experiments on challenging benchmarks demonstrate substantial improvements over existing matching pipelines, particularly under large viewpoint and scale variations. Our results show that foundation models originally designed for monocular segmentation can be effectively extended to multi-view correspondence reasoning through explicit co-visibility modeling, offering a new perspective on structured representation learning for image matching. Code and project page: https://xupan.top/Projects/samatcher

Yiyang Chen, Zhedong Zheng, Wei Ji et al.

We investigate composed image retrieval with text feedback. Users gradually look for the target of interest by moving from coarse to fine-grained feedback. However, existing methods merely focus on the latter, i.e., fine-grained search, by harnessing positive and negative pairs during training. This pair-based paradigm only considers the one-to-one distance between a pair of specific points, which is not aligned with the one-to-many coarse-grained retrieval process and compromises the recall rate. In an attempt to fill this gap, we introduce a unified learning approach to simultaneously modeling the coarse- and fine-grained retrieval by considering the multi-grained uncertainty. The key idea underpinning the proposed method is to integrate fine- and coarse-grained retrieval as matching data points with small and large fluctuations, respectively. Specifically, our method contains two modules: uncertainty modeling and uncertainty regularization. (1) The uncertainty modeling simulates the multi-grained queries by introducing identically distributed fluctuations in the feature space. (2) Based on the uncertainty modeling, we further introduce uncertainty regularization to adapt the matching objective according to the fluctuation range. Compared with existing methods, the proposed strategy explicitly prevents the model from pushing away potential candidates in the early stage, and thus improves the recall rate. On the three public datasets, i.e., FashionIQ, Fashion200k, and Shoes, the proposed method has achieved +4.03%, +3.38%, and +2.40% Recall@50 accuracy over a strong baseline, respectively.

Li Li, Wei Ji, Yiming Wu et al.

Panoptic Scene Graph Generation (PSG) parses objects and predicts their relationships (predicate) to connect human language and visual scenes. However, different language preferences of annotators and semantic overlaps between predicates lead to biased predicate annotations in the dataset, i.e. different predicates for same object pairs. Biased predicate annotations make PSG models struggle in constructing a clear decision plane among predicates, which greatly hinders the real application of PSG models. To address the intrinsic bias above, we propose a novel framework named ADTrans to adaptively transfer biased predicate annotations to informative and unified ones. To promise consistency and accuracy during the transfer process, we propose to measure the invariance of representations in each predicate class, and learn unbiased prototypes of predicates with different intensities. Meanwhile, we continuously measure the distribution changes between each presentation and its prototype, and constantly screen potential biased data. Finally, with the unbiased predicate-prototype representation embedding space, biased annotations are easily identified. Experiments show that ADTrans significantly improves the performance of benchmark models, achieving a new state-of-the-art performance, and shows great generalization and effectiveness on multiple datasets.

Shihao Zou, Yuxuan Mu, Wei Ji et al.

Event camera, as an asynchronous vision sensor capturing scene dynamics, presents new opportunities for highly efficient 3D human pose tracking. Existing approaches typically adopt modern-day Artificial Neural Networks (ANNs), such as CNNs or Transformer, where sparse events are converted into dense images or paired with additional gray-scale images as input. Such practices, however, ignore the inherent sparsity of events, resulting in redundant computations, increased energy consumption, and potentially degraded performance. Motivated by these observations, we introduce the first sparse Spiking Neural Networks (SNNs) framework for 3D human pose tracking based solely on events. Our approach eliminates the need to convert sparse data to dense formats or incorporate additional images, thereby fully exploiting the innate sparsity of input events. Central to our framework is a novel Spiking Spatiotemporal Transformer, which enables bi-directional spatiotemporal fusion of spike pose features and provides a guaranteed similarity measurement between binary spike features in spiking attention. Moreover, we have constructed a large-scale synthetic dataset, SynEventHPD, that features a broad and diverse set of 3D human motions, as well as much longer hours of event streams. Empirical experiments demonstrate the superiority of our approach over existing state-of-the-art (SOTA) ANN-based methods, requiring only 19.1% FLOPs and 3.6% energy cost. Furthermore, our approach outperforms existing SNN-based benchmarks in this task, highlighting the effectiveness of our proposed SNN framework. The dataset will be released upon acceptance, and code can be found at https://github.com/JimmyZou/HumanPoseTracking_SNN.

Jiacong Xu, Yi Zhang, Jiawei Peng et al.

Accurately estimating the 3D pose and shape is an essential step towards understanding animal behavior, and can potentially benefit many downstream applications, such as wildlife conservation. However, research in this area is held back by the lack of a comprehensive and diverse dataset with high-quality 3D pose and shape annotations. In this paper, we propose Animal3D, the first comprehensive dataset for mammal animal 3D pose and shape estimation. Animal3D consists of 3379 images collected from 40 mammal species, high-quality annotations of 26 keypoints, and importantly the pose and shape parameters of the SMAL model. All annotations were labeled and checked manually in a multi-stage process to ensure highest quality results. Based on the Animal3D dataset, we benchmark representative shape and pose estimation models at: (1) supervised learning from only the Animal3D data, (2) synthetic to real transfer from synthetically generated images, and (3) fine-tuning human pose and shape estimation models. Our experimental results demonstrate that predicting the 3D shape and pose of animals across species remains a very challenging task, despite significant advances in human pose estimation. Our results further demonstrate that synthetic pre-training is a viable strategy to boost the model performance. Overall, Animal3D opens new directions for facilitating future research in animal 3D pose and shape estimation, and is publicly available.

Qifan Yu, Juncheng Li, Yu Wu et al.

Scene Graph Generation (SGG) aims to extract <subject, predicate, object> relationships in images for vision understanding. Although recent works have made steady progress on SGG, they still suffer long-tail distribution issues that tail-predicates are more costly to train and hard to distinguish due to a small amount of annotated data compared to frequent predicates. Existing re-balancing strategies try to handle it via prior rules but are still confined to pre-defined conditions, which are not scalable for various models and datasets. In this paper, we propose a Cross-modal prediCate boosting (CaCao) framework, where a visually-prompted language model is learned to generate diverse fine-grained predicates in a low-resource way. The proposed CaCao can be applied in a plug-and-play fashion and automatically strengthen existing SGG to tackle the long-tailed problem. Based on that, we further introduce a novel Entangled cross-modal prompt approach for open-world predicate scene graph generation (Epic), where models can generalize to unseen predicates in a zero-shot manner. Comprehensive experiments on three benchmark datasets show that CaCao consistently boosts the performance of multiple scene graph generation models in a model-agnostic way. Moreover, our Epic achieves competitive performance on open-world predicate prediction. The data and code for this paper are publicly available.

Juncheng Li, Minghe Gao, Longhui Wei et al.

Prompt tuning, a recently emerging paradigm, enables the powerful vision-language pre-training models to adapt to downstream tasks in a parameter -- and data -- efficient way, by learning the ``soft prompts'' to condition frozen pre-training models. Though effective, it is particularly problematic in the few-shot scenario, where prompt tuning performance is sensitive to the initialization and requires a time-consuming process to find a good initialization, thus restricting the fast adaptation ability of the pre-training models. In addition, prompt tuning could undermine the generalizability of the pre-training models, because the learnable prompt tokens are easy to overfit to the limited training samples. To address these issues, we introduce a novel Gradient-RegulAted Meta-prompt learning (GRAM) framework that jointly meta-learns an efficient soft prompt initialization for better adaptation and a lightweight gradient regulating function for strong cross-domain generalizability in a meta-learning paradigm using only the unlabeled image-text pre-training data. Rather than designing a specific prompt tuning method, our GRAM can be easily incorporated into various prompt tuning methods in a model-agnostic way, and comprehensive experiments show that GRAM brings about consistent improvement for them in several settings (i.e., few-shot learning, cross-domain generalization, cross-dataset generalization, etc.) over 11 datasets. Further, experiments show that GRAM enables the orthogonal methods of textual and visual prompt tuning to work in a mutually-enhanced way, offering better generalizability beyond the uni-modal prompt tuning methods.

Yaoyao Zhong, Junbin Xiao, Wei Ji et al.

Video Question Answering (VideoQA) aims to answer natural language questions according to the given videos. It has earned increasing attention with recent research trends in joint vision and language understanding. Yet, compared with ImageQA, VideoQA is largely underexplored and progresses slowly. Although different algorithms have continually been proposed and shown success on different VideoQA datasets, we find that there lacks a meaningful survey to categorize them, which seriously impedes its advancements. This paper thus provides a clear taxonomy and comprehensive analyses to VideoQA, focusing on the datasets, algorithms, and unique challenges. We then point out the research trend of studying beyond factoid QA to inference QA towards the cognition of video contents, Finally, we conclude some promising directions for future exploration.

Yicong Li, Xiang Wang, Junbin Xiao et al.

Video Question Answering (VideoQA) is the task of answering questions about a video. At its core is understanding the alignments between visual scenes in video and linguistic semantics in question to yield the answer. In leading VideoQA models, the typical learning objective, empirical risk minimization (ERM), latches on superficial correlations between video-question pairs and answers as the alignments. However, ERM can be problematic, because it tends to over-exploit the spurious correlations between question-irrelevant scenes and answers, instead of inspecting the causal effect of question-critical scenes. As a result, the VideoQA models suffer from unreliable reasoning. In this work, we first take a causal look at VideoQA and argue that invariant grounding is the key to ruling out the spurious correlations. Towards this end, we propose a new learning framework, Invariant Grounding for VideoQA (IGV), to ground the question-critical scene, whose causal relations with answers are invariant across different interventions on the complement. With IGV, the VideoQA models are forced to shield the answering process from the negative influence of spurious correlations, which significantly improves the reasoning ability. Experiments on three benchmark datasets validate the superiority of IGV in terms of accuracy, visual explainability, and generalization ability over the leading baselines.

Xiang Fang, Wanlong Fang, Wei Ji et al.

Video-language models are pivotal for tasks such as moment retrieval and highlight detection, yet they often struggle to capture the dynamic, non-linear interactions between temporal video sequences and textual semantics. Existing approaches, relying on static cross-attention or prompt-tuning mechanisms, fail to adaptively model the evolving relationships between modalities, leading to suboptimal alignment and limited generalization. Inspired by systems biology, we propose \textbf{Reaction-Diffusion Multimodal Fusion (RDMF)}, a novel framework that reimagines video-language alignment as a reaction-diffusion (RD) process, drawing on the principles of pattern formation introduced by Alan Turing. In RDMF, video features diffuse across time to capture temporal context, while text-video interactions are modeled as non-linear reactions that amplify relevant features and suppress noise, forming emergent patterns akin to biological systems. Leveraging the Gray-Scott RD model, we design a computationally efficient fusion module that integrates video and text representations, supported by rigorous mathematical analysis of stability and convergence using Turing instability criteria. Our framework is theoretically grounded, employing advanced mathematical tools to ensure stable pattern formation, and is practically viable, incorporating standard components like pretrained encoders and DETR-style heads for moment retrieval and saliency prediction. RDMF represents a pioneering interdisciplinary approach, bridging systems biology and multimedia research to address the limitations of conventional multimodal fusion. Preliminary experiments demonstrate its potential to outperform existing methods in identifying salient video moments, offering a new paradigm for video-language tasks.

Ailin Huang, Chengyuan Yao, Chunrui Han et al.

We present STEP3-VL-10B, a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. STEP3-VL-10B is realized through two strategic shifts: first, a unified, fully unfrozen pre-training strategy on 1.2T multimodal tokens that integrates a language-aligned Perception Encoder with a Qwen3-8B decoder to establish intrinsic vision-language synergy; and second, a scaled post-training pipeline featuring over 1k iterations of reinforcement learning. Crucially, we implement Parallel Coordinated Reasoning (PaCoRe) to scale test-time compute, allocating resources to scalable perceptual reasoning that explores and synthesizes diverse visual hypotheses. Consequently, despite its compact 10B footprint, STEP3-VL-10B rivals or surpasses models 10$\times$-20$\times$ larger (e.g., GLM-4.6V-106B, Qwen3-VL-235B) and top-tier proprietary flagships like Gemini 2.5 Pro and Seed-1.5-VL. Delivering best-in-class performance, it records 92.2% on MMBench and 80.11% on MMMU, while excelling in complex reasoning with 94.43% on AIME2025 and 75.95% on MathVision. We release the full model suite to provide the community with a powerful, efficient, and reproducible baseline.

Yue Feng, Jingjing Li, Qijia Lu et al.

The rapid advancement of AI-driven video generation has transformed content creation, while simultaneously increasing the risk of misinformation through localized manipulations in long-form videos. Existing video forensic methods predominantly operate on short, independent clips, and thus fail to capture realistic scenarios where AI-generated content is sparsely embedded within otherwise authentic footage. To bridge this gap, we formulate the task of Temporal AI-Generated Segment Localization and Explanation, which targets authenticity detection, temporal localization, and interpretable analysis of manipulated segments in untrimmed long videos. We further introduce TASLE, a large-scale benchmark comprising 12,472 untrimmed videos with diverse manipulation patterns and rich annotation signals, including temporal boundaries, authenticity labels, and segment-level rationales. In addition, we propose MSLoc, a coarse-to-fine forensic baseline that combines a boundary-sensitive proposal generation module for efficient long-video scanning with an MLLM-based refinement module for precise boundary localization and interpretable reasoning. Experiments validate the effectiveness of the proposed baseline, highlighting the importance of segment-level explainable forensics for long-form AI-generated video analysis. Our dataset and code are publicly available at https://debby-0527.github.io/TASLE.

Wei Ji, Long Chen, Yinwei Wei et al.

Given an untrimmed video and natural language query, video sentence grounding aims to localize the target temporal moment in the video. Existing methods mainly tackle this task by matching and aligning semantics of the descriptive sentence and video segments on a single temporal resolution, while neglecting the temporal consistency of video content in different resolutions. In this work, we propose a novel multi-resolution temporal video sentence grounding network: MRTNet, which consists of a multi-modal feature encoder, a Multi-Resolution Temporal (MRT) module, and a predictor module. MRT module is an encoder-decoder network, and output features in the decoder part are in conjunction with Transformers to predict the final start and end timestamps. Particularly, our MRT module is hot-pluggable, which means it can be seamlessly incorporated into any anchor-free models. Besides, we utilize a hybrid loss to supervise cross-modal features in MRT module for more accurate grounding in three scales: frame-level, clip-level and sequence-level. Extensive experiments on three prevalent datasets have shown the effectiveness of MRTNet.

Wei Ji, Jingjing Li, Qi Bi et al.

Recently, Meta AI Research approaches a general, promptable Segment Anything Model (SAM) pre-trained on an unprecedentedly large segmentation dataset (SA-1B). Without a doubt, the emergence of SAM will yield significant benefits for a wide array of practical image segmentation applications. In this study, we conduct a series of intriguing investigations into the performance of SAM across various applications, particularly in the fields of natural images, agriculture, manufacturing, remote sensing, and healthcare. We analyze and discuss the benefits and limitations of SAM, while also presenting an outlook on its future development in segmentation tasks. By doing so, we aim to give a comprehensive understanding of SAM's practical applications. This work is expected to provide insights that facilitate future research activities toward generic segmentation. Source code is publicly available.

Hao Fei, Shengqiong Wu, Wei Ji et al.

Text-to-video (T2V) synthesis has gained increasing attention in the community, in which the recently emerged diffusion models (DMs) have promisingly shown stronger performance than the past approaches. While existing state-of-the-art DMs are competent to achieve high-resolution video generation, they may largely suffer from key limitations (e.g., action occurrence disorders, crude video motions) with respect to the intricate temporal dynamics modeling, one of the crux of video synthesis. In this work, we investigate strengthening the awareness of video dynamics for DMs, for high-quality T2V generation. Inspired by human intuition, we design an innovative dynamic scene manager (dubbed as Dysen) module, which includes (step-1) extracting from input text the key actions with proper time-order arrangement, (step-2) transforming the action schedules into the dynamic scene graph (DSG) representations, and (step-3) enriching the scenes in the DSG with sufficient and reasonable details. Taking advantage of the existing powerful LLMs (e.g., ChatGPT) via in-context learning, Dysen realizes (nearly) human-level temporal dynamics understanding. Finally, the resulting video DSG with rich action scene details is encoded as fine-grained spatio-temporal features, integrated into the backbone T2V DM for video generating. Experiments on popular T2V datasets suggest that our Dysen-VDM consistently outperforms prior arts with significant margins, especially in scenarios with complex actions. Codes at https://haofei.vip/Dysen-VDM

Jiahang Tu, Wei Ji, Hanbin Zhao et al.

In autonomous driving, deep models have shown remarkable performance across various visual perception tasks with the demand of high-quality and huge-diversity training datasets. Such datasets are expected to cover various driving scenarios with adverse weather, lighting conditions and diverse moving objects. However, manually collecting these data presents huge challenges and expensive cost. With the rapid development of large generative models, we propose DriveDiTFit, a novel method for efficiently generating autonomous Driving data by Fine-tuning pre-trained Diffusion Transformers (DiTs). Specifically, DriveDiTFit utilizes a gap-driven modulation technique to carefully select and efficiently fine-tune a few parameters in DiTs according to the discrepancy between the pre-trained source data and the target driving data. Additionally, DriveDiTFit develops an effective weather and lighting condition embedding module to ensure diversity in the generated data, which is initialized by a nearest-semantic-similarity initialization approach. Through progressive tuning scheme to refined the process of detail generation in early diffusion process and enlarging the weights corresponding to small objects in training loss, DriveDiTFit ensures high-quality generation of small moving objects in the generated data. Extensive experiments conducted on driving datasets confirm that our method could efficiently produce diverse real driving data. The source codes will be available at https://github.com/TtuHamg/DriveDiTFit.

Jingjing Li, Yue Feng, Yuchen Guo et al.

Segment Anything Model 3 (SAM3) has advanced open-vocabulary segmentation through promptable concept segmentation, allowing users to segment all instances corresponding to a given concept, typically specified with short noun-phrase (NP) prompts. While this marks the first integration of language-level concepts within the SAM family, real-world usage typically requires far richer expressions that include attributes, spatial relations, functionalities, actions, states, and even implicit reasoning over instances. Currently, SAM3 relies on external multi-modal agents to convert complex instructions into NPs and then conduct iterative mask filtering. However, these NP-level concepts remain overly coarse, often failing to precisely represent a specific instance. In this work, we present SAM3-I, an enhanced framework that unifies concept-level understanding and instruction-level reasoning within the SAM family. SAM3-I introduces an instruction-aware cascaded adaptation mechanism that progressively aligns expressive instruction semantics with SAM3's existing vision-language representations, enabling direct instruction-following segmentation without sacrificing its original concept-driven capabilities. Furthermore, we design a structured instruction taxonomy spanning concept, simple, and complex levels, and develop a scalable data engine to construct a dataset with diverse instruction-mask pairs. Experiments show that SAM3-I delivers appealing performance, demonstrating that SAM3 can be effectively extended to follow natural-language instructions while preserving its strong concept grounding. We open-source SAM3-I and provide practical fine-tuning workflows, enabling researchers to adapt it to domain-specific applications. The source code is available here.

Tianhao Peng, Haochen Wang, Yuanxing Zhang et al.

The advent of Multimodal Large Language Models (MLLMs) has expanded AI capabilities to visual modalities, yet existing evaluation benchmarks remain limited to single-video understanding, overlooking the critical need for multi-video understanding in real-world scenarios (e.g., sports analytics and autonomous driving). To address this significant gap, we introduce MVU-Eval, the first comprehensive benchmark for evaluating Multi-Video Understanding for MLLMs. Specifically, our MVU-Eval mainly assesses eight core competencies through 1,824 meticulously curated question-answer pairs spanning 4,959 videos from diverse domains, addressing both fundamental perception tasks and high-order reasoning tasks. These capabilities are rigorously aligned with real-world applications such as multi-sensor synthesis in autonomous systems and cross-angle sports analytics. Through extensive evaluation of state-of-the-art open-source and closed-source models, we reveal significant performance discrepancies and limitations in current MLLMs' ability to perform understanding across multiple videos. The benchmark will be made publicly available to foster future research.

Kai Peng, Yunzhe Shen, Miao Zhang et al.

The ability to capture and segment sounding objects in dynamic visual scenes is crucial for the development of Audio-Visual Segmentation (AVS) tasks. While significant progress has been made in this area, the interaction between audio and visual modalities still requires further exploration. In this work, we aim to answer the following questions: How can a model effectively suppress audio noise while enhancing relevant audio information? How can we achieve discriminative interaction between the audio and visual modalities? To this end, we propose SDAVS, equipped with the Selective Noise-Resilient Processor (SNRP) module and the Discriminative Audio-Visual Mutual Fusion (DAMF) strategy. The proposed SNRP mitigates audio noise interference by selectively emphasizing relevant auditory cues, while DAMF ensures more consistent audio-visual representations. Experimental results demonstrate that our proposed method achieves state-of-the-art performance on benchmark AVS datasets, especially in multi-source and complex scenes. \textit{The code and model are available at https://github.com/happylife-pk/SDAVS}.

Wei Ji, Jingjing Li, Qi Bi et al.

Growing interests in RGB-D salient object detection (RGB-D SOD) have been witnessed in recent years, owing partly to the popularity of depth sensors and the rapid progress of deep learning techniques. Unfortunately, existing RGB-D SOD methods typically demand large quantity of training images being thoroughly annotated at pixel-level. The laborious and time-consuming manual annotation has become a real bottleneck in various practical scenarios. On the other hand, current unsupervised RGB-D SOD methods still heavily rely on handcrafted feature representations. This inspires us to propose in this paper a deep unsupervised RGB-D saliency detection approach, which requires no manual pixel-level annotation during training. It is realized by two key ingredients in our training pipeline. First, a depth-disentangled saliency update (DSU) framework is designed to automatically produce pseudo-labels with iterative follow-up refinements, which provides more trustworthy supervision signals for training the saliency network. Second, an attentive training strategy is introduced to tackle the issue of noisy pseudo-labels, by properly re-weighting to highlight the more reliable pseudo-labels. Extensive experiments demonstrate the superior efficiency and effectiveness of our approach in tackling the challenging unsupervised RGB-D SOD scenarios. Moreover, our approach can also be adapted to work in fully-supervised situation. Empirical studies show the incorporation of our approach gives rise to notably performance improvement in existing supervised RGB-D SOD models.

Tao Wu, Mengze Li, Jingyuan Chen et al.

Research on Multi-modal Large Language Models (MLLMs) towards the multi-image cross-modal instruction has received increasing attention and made significant progress, particularly in scenarios involving closely resembling images (e.g., change captioning). Existing MLLMs typically follow a two-step process in their pipelines: first, extracting visual tokens independently for each input image, and then aligning these visual tokens from different images with the Large Language Model (LLM) in its textual feature space. However, the independent extraction of visual tokens for each image may result in different semantics being prioritized for different images in the first step, leading to a lack of preservation of linking information among images for subsequent LLM analysis. This issue becomes more serious in scenarios where significant variations exist among the images (e.g., visual storytelling). To address this challenge, we introduce Semantic Alignment for Multi-modal large language models (SAM). By involving the bidirectional semantic guidance between different images in the visual-token extraction process, SAM aims to enhance the preservation of linking information for coherent analysis and align the semantics of different images before feeding them into LLM. As the test bed, we propose a large-scale dataset named MmLINK consisting of 69K samples. Different from most existing datasets for MLLMs fine-tuning, our MmLINK dataset comprises multi-modal instructions with significantly diverse images. Extensive experiments on the group captioning task and the storytelling task prove the effectiveness of our SAM model, surpassing the state-of-the-art methods by a large margin (+37% for group captioning and +22% for storytelling on CIDEr score). Project page: https://mccartney01.github.io/SAM.

Kaihang Pan, Juncheng Li, Wenjie Wang et al.

Recent studies indicate that dense retrieval models struggle to perform well on a wide variety of retrieval tasks that lack dedicated training data, as different retrieval tasks often entail distinct search intents. To address this challenge, in this work we leverage instructions to flexibly describe retrieval intents and introduce I3, a unified retrieval system that performs Intent-Introspective retrieval across various tasks, conditioned on Instructions without any task-specific training. I3 innovatively incorporates a pluggable introspector in a parameter-isolated manner to comprehend specific retrieval intents by jointly reasoning over the input query and instruction, and seamlessly integrates the introspected intent into the original retrieval model for intent-aware retrieval. Furthermore, we propose progressively-pruned intent learning. It utilizes extensive LLM-generated data to train I3 phase-by-phase, embodying two key designs: progressive structure pruning and drawback extrapolation-based data refinement. Extensive experiments show that in the BEIR benchmark, I3 significantly outperforms baseline methods designed with task-specific retrievers, achieving state-of-the-art zero-shot performance without any task-specific tuning.

Meng Wei, Long Chen, Wei Ji et al.

Video Visual Relation Detection (VidVRD) aims to detect visual relationship triplets in videos using spatial bounding boxes and temporal boundaries. Existing VidVRD methods can be broadly categorized into bottom-up and top-down paradigms, depending on their approach to classifying relations. Bottom-up methods follow a clip-based approach where they classify relations of short clip tubelet pairs and then merge them into long video relations. On the other hand, top-down methods directly classify long video tubelet pairs. While recent video-based methods utilizing video tubelets have shown promising results, we argue that the effective modeling of spatial and temporal context plays a more significant role than the choice between clip tubelets and video tubelets. This motivates us to revisit the clip-based paradigm and explore the key success factors in VidVRD. In this paper, we propose a Hierarchical Context Model (HCM) that enriches the object-based spatial context and relation-based temporal context based on clips. We demonstrate that using clip tubelets can achieve superior performance compared to most video-based methods. Additionally, using clip tubelets offers more flexibility in model designs and helps alleviate the limitations associated with video tubelets, such as the challenging long-term object tracking problem and the loss of temporal information in long-term tubelet feature compression. Extensive experiments conducted on two challenging VidVRD benchmarks validate that our HCM achieves a new state-of-the-art performance, highlighting the effectiveness of incorporating advanced spatial and temporal context modeling within the clip-based paradigm.

Minghe Gao, Juncheng Li, Hao Fei et al.

Visual programming, a modular and generalizable paradigm, integrates different modules and Python operators to solve various vision-language tasks. Unlike end-to-end models that need task-specific data, it advances in performing visual processing and reasoning in an unsupervised manner. Current visual programming methods generate programs in a single pass for each task where the ability to evaluate and optimize based on feedback, unfortunately, is lacking, which consequentially limits their effectiveness for complex, multi-step problems. Drawing inspiration from benders decomposition, we introduce De-fine, a training-free framework that automatically decomposes complex tasks into simpler subtasks and refines programs through auto-feedback. This model-agnostic approach can improve logical reasoning performance by integrating the strengths of multiple models. Our experiments across various visual tasks show that De-fine creates more robust programs. Moreover, viewing each feedback module as an independent agent will yield fresh prospects for the field of agent research.

Jingjun Yi, Qi Bi, Hao Zheng et al.

The rapid development of Vision Foundation Model (VFM) brings inherent out-domain generalization for a variety of down-stream tasks. Among them, domain generalized semantic segmentation (DGSS) holds unique challenges as the cross-domain images share common pixel-wise content information but vary greatly in terms of the style. In this paper, we present a novel Spectral-dEcomposed Token (SET) learning framework to advance the frontier. Delving into further than existing fine-tuning token & frozen backbone paradigm, the proposed SET especially focuses on the way learning style-invariant features from these learnable tokens. Particularly, the frozen VFM features are first decomposed into the phase and amplitude components in the frequency space, which mainly contain the information of content and style, respectively, and then separately processed by learnable tokens for task-specific information extraction. After the decomposition, style variation primarily impacts the token-based feature enhancement within the amplitude branch. To address this issue, we further develop an attention optimization method to bridge the gap between style-affected representation and static tokens during inference. Extensive cross-domain experiments show its state-of-the-art performance.

Meng Chu, Zhedong Zheng, Wei Ji et al.

Navigating drones through natural language commands remains challenging due to the dearth of accessible multi-modal datasets and the stringent precision requirements for aligning visual and textual data. To address this pressing need, we introduce GeoText-1652, a new natural language-guided geo-localization benchmark. This dataset is systematically constructed through an interactive human-computer process leveraging Large Language Model (LLM) driven annotation techniques in conjunction with pre-trained vision models. GeoText-1652 extends the established University-1652 image dataset with spatial-aware text annotations, thereby establishing one-to-one correspondences between image, text, and bounding box elements. We further introduce a new optimization objective to leverage fine-grained spatial associations, called blending spatial matching, for region-level spatial relation matching. Extensive experiments reveal that our approach maintains a competitive recall rate comparing other prevailing cross-modality methods. This underscores the promising potential of our approach in elevating drone control and navigation through the seamless integration of natural language commands in real-world scenarios.

Ailin Huang, Ang Li, Aobo Kong et al.

We introduce Step 3.5 Flash, a sparse Mixture-of-Experts (MoE) model that bridges frontier-level agentic intelligence and computational efficiency. We focus on what matters most when building agents: sharp reasoning and fast, reliable execution. Step 3.5 Flash pairs a 196B-parameter foundation with 11B active parameters for efficient inference. It is optimized with interleaved 3:1 sliding-window/full attention and Multi-Token Prediction (MTP-3) to reduce the latency and cost of multi-round agentic interactions. To reach frontier-level intelligence, we design a scalable reinforcement learning framework that combines verifiable signals with preference feedback, while remaining stable under large-scale off-policy training, enabling consistent self-improvement across mathematics, code, and tool use. Step 3.5 Flash demonstrates strong performance across agent, coding, and math tasks, achieving 85.4% on IMO-AnswerBench, 86.4% on LiveCodeBench-v6 (2024.08-2025.05), 88.2% on tau2-Bench, 69.0% on BrowseComp (with context management), and 51.0% on Terminal-Bench 2.0, comparable to frontier models such as GPT-5.2 xHigh and Gemini 3.0 Pro. By redefining the efficiency frontier, Step 3.5 Flash provides a high-density foundation for deploying sophisticated agents in real-world industrial environments.

Yang Chen, Shanshan Zhao, Wei Ji et al.

Relying on deep supervised or self-supervised learning, previous methods for depth completion from paired single image and sparse depth data have achieved impressive performance in recent years. However, facing a new environment where the test data occurs online and differs from the training data in the RGB image content and depth sparsity, the trained model might suffer severe performance drop. To encourage the trained model to work well in such conditions, we expect it to be capable of adapting to the new environment continuously and effectively. To achieve this, we propose MetaComp. It utilizes the meta-learning technique to simulate adaptation policies during the training phase, and then adapts the model to new environments in a self-supervised manner in testing. Considering that the input is multi-modal data, it would be challenging to adapt a model to variations in two modalities simultaneously, due to significant differences in structure and form of the two modal data. Therefore, we further propose to disentangle the adaptation procedure in the basic meta-learning training into two steps, the first one focusing on the depth sparsity while the second attending to the image content. During testing, we take the same strategy to adapt the model online to new multi-modal data. Experimental results and comprehensive ablations show that our MetaComp is capable of adapting to the depth completion in a new environment effectively and robust to changes in different modalities.

Wei Ji, Li Li, Hao Fei et al.

Referring Image Understanding (RIS) has been extensively studied over the past decade, leading to the development of advanced algorithms. However, there has been a lack of research investigating how existing algorithms should be benchmarked with complex language queries, which include more informative descriptions of surrounding objects and backgrounds (\eg \textit{"the black car."} vs. \textit{"the black car is parking on the road and beside the bus."}). Given the significant improvement in the semantic understanding capability of large pre-trained models, it is crucial to take a step further in RIS by incorporating complex language that resembles real-world applications. To close this gap, building upon the existing RefCOCO and Visual Genome datasets, we propose a new RIS benchmark with complex queries, namely \textbf{RIS-CQ}. The RIS-CQ dataset is of high quality and large scale, which challenges the existing RIS with enriched, specific and informative queries, and enables a more realistic scenario of RIS research. Besides, we present a nichetargeting method to better task the RIS-CQ, called dual-modality graph alignment model (\textbf{\textsc{DuMoGa}}), which outperforms a series of RIS methods.

Xiangyan Liu, Rongxue Li, Wei Ji et al.

The reasoning capabilities of LLM (Large Language Model) are widely acknowledged in recent research, inspiring studies on tool learning and autonomous agents. LLM serves as the "brain" of the agent, orchestrating multiple tools for collaborative multi-step task solving. Unlike methods invoking tools like calculators or weather APIs for straightforward tasks, multi-modal agents excel by integrating diverse AI models for complex challenges. However, current multi-modal agents neglect the significance of model selection: they primarily focus on the planning and execution phases, and will only invoke predefined task-specific models for each subtask, making the execution fragile. Meanwhile, other traditional model selection methods are either incompatible with or suboptimal for the multi-modal agent scenarios, due to ignorance of dependencies among subtasks arising by multi-step reasoning. To this end, we identify the key challenges therein and propose the $\textit{M}^3$ framework as a plug-in with negligible runtime overhead at test-time. This framework improves model selection and bolsters the robustness of multi-modal agents in multi-step reasoning. In the absence of suitable benchmarks, we create MS-GQA, a new dataset specifically designed to investigate the model selection challenge in multi-modal agents. Our experiments reveal that our framework enables dynamic model selection, considering both user inputs and subtask dependencies, thereby robustifying the overall reasoning process. Our code and benchmark: https://github.com/LINs-lab/M3.

Xiaolu Liu, Yicong Li, Qiyuan He et al.

Textured 3D morphing seeks to generate smooth and plausible transitions between two 3D assets, preserving both structural coherence and fine-grained appearance. This ability is crucial not only for advancing 3D generation research but also for practical applications in animation, editing, and digital content creation. Existing approaches either operate directly on geometry, limiting them to shape-only morphing while neglecting textures, or extend 2D interpolation strategies into 3D, which often causes semantic ambiguity, structural misalignment, and texture blurring. These challenges underscore the necessity to jointly preserve geometric consistency, texture alignment, and robustness throughout the transition process. To address this, we propose Interp3D, a novel training-free framework for textured 3D morphing. It harnesses generative priors and adopts a progressive alignment principle to ensure both geometric fidelity and texture coherence. Starting from semantically aligned interpolation in condition space, Interp3D enforces structural consistency via SLAT (Structured Latent)-guided structure interpolation, and finally transfers appearance details through fine-grained texture fusion. For comprehensive evaluations, we construct a dedicated dataset, Interp3DData, with graded difficulty levels and assess generation results from fidelity, transition smoothness, and plausibility. Both quantitative metrics and human studies demonstrate the significant advantages of our proposed approach over previous methods. Source code is available at https://github.com/xiaolul2/Interp3D.

Zhen Zhang, Wanjing Zhou, Juncheng Li et al.

Compared with individual agents, large language model based multi-agent systems have shown great capabilities consistently across diverse tasks, including code generation, mathematical reasoning, and planning, etc. Despite their impressive performance, the effectiveness and robustness of these systems heavily rely on their communication topology, which is often fixed or generated in a single step. This restricts fine-grained structural exploration and flexible composition, resulting in excessive token utilization on simple tasks while limiting capability on complicated tasks. To mitigate this challenge, we introduce RADAR, a redundancy-aware and query-adaptive generative framework that actively reduce communication overhead. Motivated by recent progress in conditional discrete graph diffusion models, we formulate communication topology design as a step-by-step generation process, guided by the effective size of the graph. Comprehensive experiments on six benchmarks demonstrate that RADAR consistently outperforms recent baselines, achieving higher accuracy, lower token consumption, and greater robustness across diverse scenarios. Our code and data are available at https://github.com/cszhangzhen/RADAR.

Guoqing Ma, Haoyang Huang, Kun Yan et al.

We present Step-Video-T2V, a state-of-the-art text-to-video pre-trained model with 30B parameters and the ability to generate videos up to 204 frames in length. A deep compression Variational Autoencoder, Video-VAE, is designed for video generation tasks, achieving 16x16 spatial and 8x temporal compression ratios, while maintaining exceptional video reconstruction quality. User prompts are encoded using two bilingual text encoders to handle both English and Chinese. A DiT with 3D full attention is trained using Flow Matching and is employed to denoise input noise into latent frames. A video-based DPO approach, Video-DPO, is applied to reduce artifacts and improve the visual quality of the generated videos. We also detail our training strategies and share key observations and insights. Step-Video-T2V's performance is evaluated on a novel video generation benchmark, Step-Video-T2V-Eval, demonstrating its state-of-the-art text-to-video quality when compared with both open-source and commercial engines. Additionally, we discuss the limitations of current diffusion-based model paradigm and outline future directions for video foundation models. We make both Step-Video-T2V and Step-Video-T2V-Eval available at https://github.com/stepfun-ai/Step-Video-T2V. The online version can be accessed from https://yuewen.cn/videos as well. Our goal is to accelerate the innovation of video foundation models and empower video content creators.

Bo Zhao, Yihang Liu, Chenfeng Zhang et al.

Text-guided texture editing aims to modify object appearance while preserving the underlying geometric structure. However, our empirical analysis reveals that even SOTA editing models frequently struggle to maintain structural consistency during texture editing, despite the intended changes being purely appearance-related. Motivated by this observation, we jointly enhance structure preservation from both data and training perspectives, and build TexEditor, a dedicated texture editing model based on Qwen-Image-Edit-2509. Firstly, we construct TexBlender, a high-quality SFT dataset generated with Blender, which provides strong structural priors for a cold start. Sec- ondly, we introduce StructureNFT, a RL-based approach that integrates structure-preserving losses to transfer the structural priors learned during SFT to real-world scenes. Moreover, due to the limited realism and evaluation coverage of existing benchmarks, we introduce TexBench, a general-purpose real-world benchmark for text-guided texture editing. Extensive experiments on existing Blender-based texture benchmarks and our TexBench show that TexEditor consistently outperforms strong baselines such as Nano Banana Pro. In addition, we assess TexEditor on the general purpose benchmark ImgEdit to validate its generalization. Our code and data are available at https://github.com/KlingAIResearch/TexEditor.

Ailin Huang, Boyong Wu, Bruce Wang et al.

Real-time speech interaction, serving as a fundamental interface for human-machine collaboration, holds immense potential. However, current open-source models face limitations such as high costs in voice data collection, weakness in dynamic control, and limited intelligence. To address these challenges, this paper introduces Step-Audio, the first production-ready open-source solution. Key contributions include: 1) a 130B-parameter unified speech-text multi-modal model that achieves unified understanding and generation, with the Step-Audio-Chat version open-sourced; 2) a generative speech data engine that establishes an affordable voice cloning framework and produces the open-sourced lightweight Step-Audio-TTS-3B model through distillation; 3) an instruction-driven fine control system enabling dynamic adjustments across dialects, emotions, singing, and RAP; 4) an enhanced cognitive architecture augmented with tool calling and role-playing abilities to manage complex tasks effectively. Based on our new StepEval-Audio-360 evaluation benchmark, Step-Audio achieves state-of-the-art performance in human evaluations, especially in terms of instruction following. On open-source benchmarks like LLaMA Question, shows 9.3% average performance improvement, demonstrating our commitment to advancing the development of open-source multi-modal language technologies. Our code and models are available at https://github.com/stepfun-ai/Step-Audio.

Li Li, You Qin, Wei Ji et al.

Panoptic Scene Graph Generation (PSG) involves the detection of objects and the prediction of their corresponding relationships (predicates). However, the presence of biased predicate annotations poses a significant challenge for PSG models, as it hinders their ability to establish a clear decision boundary among different predicates. This issue substantially impedes the practical utility and real-world applicability of PSG models. To address the intrinsic bias above, we propose a novel framework to infer potentially biased annotations by measuring the predicate prediction risks within each subject-object pair (domain), and adaptively transfer the biased annotations to consistent ones by learning invariant predicate representation embeddings. Experiments show that our method significantly improves the performance of benchmark models, achieving a new state-of-the-art performance, and shows great generalization and effectiveness on PSG dataset.

Zhongyi Guo, Keji Han, Yao Ge et al.

Deep neural networks (DNNs) can be easily fooled by adversarial attacks during inference phase when attackers add imperceptible perturbations to original examples, i.e., adversarial examples. Many works focus on adversarial detection and adversarial training to defend against adversarial attacks. However, few works explore the tool-chains behind adversarial examples, which can help defenders to seize the clues about the originator of the attack, their goals, and provide insight into the most effective defense algorithm against corresponding attacks. With such a gap, it is necessary to develop techniques that can recognize tool-chains that are leveraged to generate the adversarial examples, which is called Adversarial Attribution Problem (AAP). In this paper, AAP is defined as the recognition of three signatures, i.e., {\em attack algorithm}, {\em victim model} and {\em hyperparameter}. Current works transfer AAP into single label classification task and ignore the relationship between these signatures. The former will meet combination explosion problem as the number of signatures is increasing. The latter dictates that we cannot treat AAP simply as a single task problem. We first conduct some experiments to validate the attributability of adversarial examples. Furthermore, we propose a multi-task learning framework named Multi-Task Adversarial Attribution (MTAA) to recognize the three signatures simultaneously. MTAA contains perturbation extraction module, adversarial-only extraction module and classification and regression module. It takes the relationship between attack algorithm and corresponding hyperparameter into account and uses the uncertainty weighted loss to adjust the weights of three recognition tasks. The experimental results on MNIST and ImageNet show the feasibility and scalability of the proposed framework as well as its effectiveness in dealing with false alarms.

Boyong Wu, Chao Yan, Chen Hu et al.

This paper presents Step-Audio 2, an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation. By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding. To facilitate genuine end-to-end speech conversation, Step-Audio 2 incorporates the generation of discrete audio tokens into language modeling, significantly enhancing its responsiveness to paralinguistic information such as speaking styles and emotions. To effectively leverage the rich textual and acoustic knowledge in real-world data, Step-Audio 2 integrates retrieval-augmented generation (RAG) and is able to call external tools such as web search to mitigate hallucination and audio search to switch timbres. Trained on millions of hours of speech and audio data, Step-Audio 2 delivers intelligence and expressiveness across diverse conversational scenarios. Evaluation results demonstrate that Step-Audio 2 achieves state-of-the-art performance on various audio understanding and conversational benchmarks compared to other open-source and commercial solutions. Please visit https://github.com/stepfun-ai/Step-Audio2 for more information.