论文检索

Ying Sun, Wenjun Wang, Nannan Wu, Chunlong Bao

Anomaly subgraph detection is a crucial task in various real-world applications, including identifying high-risk areas, detecting river pollution, and monitoring disease outbreaks. Early traditional graph-based methods can obtain high-precision detection results in scenes with small-scale graphs and obvious anomaly features. Most existing anomaly detection methods based on deep learning primarily concentrate on identifying anomalies at the node level, while neglecting to detect anomaly groups in the internal structure. In this paper, we propose a novel end-to-end Graph Neural Network (GNN) based anomaly subgraph detection approach(ASD-HC) in graph-structured data. 1)We propose a high-order neighborhood sampling strategy to construct our node and k-order neighbor-subgraph instance pairs. 2)Anomaly features of nodes are captured through a self-supervised contrastive learning model. 3) Detecting the maximum connected anomaly subgraph is performed by integrating the Non-parameter Graph Scan statistics and a Random Walk module. We evaluate ASD-HC against five state-of-the-art baselines using five benchmark datasets. ASD-HC outperforms the baselines by over 13.01% in AUC score. Various experiments demonstrate that our approach effectively detects anomaly subgraphs within large-scale graphs.

Zhengdao Li, Yong Cao, Kefan Shuai, Yiming Miao, Kai Hwang

Graph classification benchmarks, vital for assessing and developing graph neural network (GNN) models, have recently been scrutinized, as simple methods like MLPs have demonstrated comparable performance. This leads to an important question: Do these benchmarks effectively distinguish the advancements of GNNs over other methodologies? If so, how do we quantitatively measure this effectiveness? In response, we first propose an empirical protocol based on a fair benchmarking framework to investigate the performance discrepancy between simple methods and GNNs. We further propose a novel metric to quantify the dataset effectiveness by considering both dataset complexity and model performance. To the best of our knowledge, our work is the first to thoroughly study and provide an explicit definition for dataset effectiveness in the graph learning area. Through testing across 16 real-world datasets, we found our metric to align with existing studies and intuitive assumptions. Finally, we explore the causes behind the low effectiveness of certain datasets by investigating the correlation between intrinsic graph properties and class labels, and we developed a novel technique supporting the correlation-controllable synthetic dataset generation. Our findings shed light on the current understanding of benchmark datasets, and our new platform could fuel the future evolution of graph classification benchmarks.

Leilei Ding, Dazhong Shen, Chao Wang, Tianfu Wang, Le Zhang, Yanyong Zhang

Graph Convolutional Networks (GCNs) have become pivotal in recommendation systems for learning user and item embeddings by leveraging the user-item interaction graph's node information and topology. However, these models often face the famous over-smoothing issue, leading to indistinct user and item embeddings and reduced personalization. Traditional desmoothing methods in GCN-based systems are model-specific, lacking a universal solution. This paper introduces a novel, model-agnostic approach named Desmoothing Framework for GCN-based Recommendation Systems (DGR). It effectively addresses over-smoothing on general GCN-based recommendation models by considering both global and local perspectives. Specifically, we first introduce vector perturbations during each message passing layer to penalize the tendency of node embeddings approximating overly to be similar with the guidance of the global topological structure. Meanwhile, we further develop a tailored-design loss term for the readout embeddings to preserve the local collaborative relations between users and their neighboring items. In particular, items that exhibit a high correlation with neighboring items are also incorporated to enhance the local topological information. To validate our approach, we conduct extensive experiments on 5 benchmark datasets based on 5 well-known GCN-based recommendation models, demonstrating the effectiveness and generalization of our proposed framework. Our code is available at https://github.com/me-sonandme/DGR.

Peng Zhao, Yin Wang, Wei Wang, Jie Mu, Huiting Liu, Cong Wang, Xiaochun Cao

Few-Shot Learning (FSL) aims to train a model that can generalize to recognize new classes, with each new class having only very limited training samples. Since extracting discriminative features for new classes with few samples is challenging, existing FSL methods leverage visual and semantic prior knowledge to guide discriminative feature learning. However, for meta-learning purposes, the semantic knowledge of the query set is unavailable, so their features lack discriminability. To address this problem, we propose a novel Multi-Attention based Visual-Semantic Interaction (MAVSI) approach for FSL. Specifically, we utilize spatial and channel attention mechanisms to effectively select discriminative visual features for the support set based on its ground-truth semantics while using all the support set semantics for each query set sample. Then, a relation module with class prototypes of the support set is employed to supervise and select discriminative visual features for the query set. To further enhance the discriminability of the support set, we introduce a visual-semantic contrastive learning module to promote the similarity between visual features and their corresponding semantic features. Extensive experiments on four benchmark datasets demonstrate that our proposed MAVSI could outperform existing state-of-the-art FSL methods.

Junjie Zhang, Tianci Hu, Xiaoshui Huang, Yongshun Gong, Dan Zeng

Evaluating the performance of Multi-modal Large Language Models (MLLMs), integrating both point cloud and language, presents significant challenges. The lack of a comprehensive assessment hampers determining whether these models truly represent advancements, thereby impeding further progress in the field. Current evaluations heavily rely on classification and caption tasks, falling short in providing a thorough assessment of MLLMs. A pressing need exists for a more sophisticated evaluation method capable of thoroughly analyzing the spatial understanding and expressive capabilities of these models. To address these issues, we introduce a scalable 3D benchmark, accompanied by a large-scale instruction-tuning dataset known as 3DBench, providing an extensible platform for a comprehensive evaluation of MLLMs. Specifically, we establish the benchmark that spans a wide range of spatial and semantic scales, from object-level to scene-level, addressing both perception and planning tasks. Furthermore, we present a rigorous pipeline for automatically constructing scalable 3D instruction-tuning datasets, covering 10 diverse multi-modal tasks with more than 0.23 million QA pairs generated in total. Thorough experiments evaluating trending MLLMs, comparisons against existing datasets, and variations of training protocols demonstrate the superiority of 3DBench, offering valuable insights into current limitations and potential research directions. Codes are available at https://github.com/Inshsang/3DBench.

Mengmeng Zhan, Zongqian Wu, Rongyao Hu, Ping Hu, Heng Tao Shen, Xiaofeng Zhu

In domain adaptation, challenges such as data privacy constraints can impede access to source data, catalyzing the development of source-free domain adaptation (SFDA) methods. However, current approaches heavily rely on models trained on source data, posing the risk of overfitting and suboptimal generalization.This paper introduces a dynamic prompt learning paradigm that harnesses the power of large-scale vision-language models to enhance the semantic transfer of source models. Specifically, our approach fosters robust and adaptive collaboration between the source-trained model and the vision-language model, facilitating the reliable extraction of domain-specific information from unlabeled target data, while consolidating domain-invariant knowledge. Without the need for accessing source data, our method amalgamates the strengths inherent in both traditional SFDA approaches and vision-language models, formulating a collaborative framework for addressing SFDA challenges. Extensive experiments conducted on three benchmark datasets showcase the superiority of our framework over previous SOTA methods.

Huibin Yan, Shuoyao Wang

The U-shape (encoder-decoder) architecture, combined with effective blocks, has shown significant success in image restoration. In U-shape models, there is insufficient focus on the feature fusion problem between encoder and decoder features at the same level. Current methods often employ simplistic operations like summation or concatenation, which makes it difficult to strike a balance between performance and complexity. To address this issue, we propose a compression-in-the-middle mechanism, termed Integration-Compression-Integration (ICI), which effectively conducts dense fusion and avoids information loss. From the block design perspective, we design a multi-dimension aggregation (MDA) mechanism, capable of effectively aggregating features from both the channel and spatial dimension. Combining the IntegrationCompression-Integration feature fusion and the multi-dimension aggregation, our dense fusion and multi-dimension aggregation network (DFMDANet) achieves superior performance over state-ofthe-art algorithms on 16 benchmarking datasets for numerous image restoration tasks.

Yang Xu, Yifan Feng, Yue Gao

The limited availability of supervised labels (positive information) poses a notable challenge for open-world retrieval. However, negative information is more easily obtained but remains underexploited in current methods. In this paper, we introduce the Negative Prompt Driven Complementary Parallel Representation (NPCP) framework, which navigates the complexities of open-world retrieval through the lens of Negative Prompts. Specifically, we employ the Parallel Exclusive Embedding (PEE) to effectively utilize the prompt information, bilaterally capturing both explicit negative and implicit positive signals. To address the challenges of embedding unification and generalization, our method leverages high-order correlations among objects through the Complementary Structure Tuning (CST), by constructing a complementary hypergraph based on bi-directional and cross-category correlations. We have developed four multimodal datasets for open-world 3D object retrieval with negative prompts: NPMN, NPAB, NPNT, and NPES. Extensive experiments and ablation studies on these four benchmarks demonstrate the superiority of our method over current state-of-the-art approaches.

Wei Wang, Hanyang Li, Ke Shi, Chao Huang, Yang Cao, Cong Wang, Xiaochun Cao

Domain adaptation aims to achieve label transfer from a labeled source domain to an unlabeled target domain, where the two domains exhibit different distributions. Existing methods primarily concentrate on designing a feature extractor to learn better domain-invariant features, along with developing an effective classifier for reliable predictions. In this paper, we introduce optimal graph learning to generate a cross-domain graph that effectively connects the two domains, and two domain-specific graphs to capture domain-specific structures. On the one hand, we incorporate the three graphs into the label propagation (LP) classifier to enhance its robustness to distribution difference. On the other hand, we leverage the three graphs to introduce graph embedding losses, promoting the learning of locally discriminative and domain-invariant features. Furthermore, we maximize the nuclear norm of predictions in LP to enhance class diversity, thereby improving its robustness to class imbalance problem. Correspondingly, we develop an efficient algorithm to solve the associated optimization problem. Finally, we integrate the proposed LP and graph embedding losses into a deep neural network, resulting in our proposed deep cross-domain robust LP. Extensive experiments conducted on three cross-domain benchmark datasets demonstrate that our proposed approach could outperform existing state-of-the-art domain adaptation methods.

Chenhui Wang, Tao Chen, Zhihao Chen, Zhizhong Huang, Taoran Jiang, Qi Wang, Hongming Shan

Despite their impressive generative performance, latent diffusion model-based virtual try-on (VTON) methods lack faithfulness to crucial details of the clothes, such as style, pattern, and text. To alleviate these issues caused by the diffusion stochastic nature and latent supervision, we propose a novel Faithful Latent Diffusion Model for VTON, termed FLDM-VTON. FLDM-VTON improves the conventional latent diffusion process in three major aspects. First, we propose incorporating warped clothes as both the starting point and local condition, supplying the model with faithful clothes priors. Second, we introduce a novel clothes flattening network to constrain generated try-on images, providing clothes-consistent faithful supervision. Third, we devise a clothes-posterior sampling for faithful inference, further enhancing the model performance over conventional clothes-agnostic Gaussian sampling. Extensive experimental results on the benchmark VITON-HD and Dress Code datasets demonstrate that our FLDM-VTON outperforms state-of-the-art baselines and is able to generate photo-realistic try-on images with faithful clothing details.

Bo Wang, Zhao Zhang, Mingbo Zhao, Xiaojie Jin, Mingliang Xu, Meng Wang

Mainstream image captioning models are usually two-stage captioners, i.e., encoding the region features by a pre-trained detector and then feeding them into a language model to generate the captions. However, such a two-stage procedure will lead to a task-based information gap that decreases the performance, because the region features in the detection task are suboptimal representations and cannot provide all the necessary information for subsequent captions generation. Besides, the region features are usually represented from the last layer of the detectors that lose the local details of images. In this paper, we propose a novel One-Stage Image Captioner (OSIC) with dynamic multi-sight learning, which directly transforms the images into descriptive sentences in one stage for eliminating the information gap. Specifically, to obtain rich features, multi-level features are captured by Swin Transformer, and then fed into a novel dynamic multi-sight embedding module to exploit both the global structure and local texture of input images. To enhance the global modeling capacity of the visual encoder, we propose a new dual-dimensional refining to non-locally model the features interaction. As a result, OSIC can directly obtain rich semantic information to improve the captioner. Extensive comparisons on the benchmark MS-COCO, Flickr8K and Flickr30K datasets verified the superior performance of our method.

Hongbo Sun, Jiahuan Zhou, Xiangteng He, Jinglin Xu, Yuxin Peng

Few-Shot Class Incremental Learning (FSCIL) aims to continually learn new classes with few training samples without forgetting already learned old classes. Existing FSCIL methods generally fix the backbone network in incremental sessions to achieve a balance between suppressing forgetting old classes and learning new classes. However, the fixed backbone network causes insufficient learning of new classes from a few samples. Benefiting from the powerful visual and textual understanding ability of Vision-Language (VL) pre-training models, we propose a Fine-grained Feature Mining Prompt Learning (FineFMPL) approach to adapt the VL model to FSCIL, which comprehensively learns and memorizes fine-grained discriminative information of emerging classes. Concretely, the visual probe prompt is firstly proposed to guide the image encoder of VL model to extract global-level coarse-grained features and object-level fine-grained features, and visual prototypes are preserved based on image patch significance, which contains the discriminative characteristics exclusive to the class. Secondly, the textual context prompt is constructed by cross-modal mapping of visual prototypes, feeding into the text encoder of VL model to memorize the class information as textual prototypes. Finally, integrating visual and textual prototypes based on fine-grained feature mining into the model improves the recognition performance of all classes in FSCIL. Extensive experiments on three benchmark datasets demonstrate that our FineFMPL achieves new state-of-the-art. The code is available at https://github.com/PKU-ICST-MIPL/FineFMPL_IJCAI2024.

Ziying Song, Guoxing Zhang, Lin Liu, Lei Yang, Shaoqing Xu, Caiyan Jia, Feiyang Jia, Li Wang

Multi-modal 3D object detectors are dedicated to exploring secure and reliable perception systems for autonomous driving (AD). Although achieving state-of-the-art (SOTA) performance on clean benchmark datasets, they tend to overlook the complexity and harsh conditions of real-world environments. With the emergence of visual foundation models (VFMs), opportunities and challenges are presented for improving the robustness and generalization of multi-modal 3D object detection in AD. Therefore, we propose RoboFusion, a robust framework that leverages VFMs like SAM to tackle out-of-distribution (OOD) noise scenarios. We first adapt the original SAM for AD scenarios named SAM-AD. To align SAM or SAM-AD with multi-modal methods, we then introduce AD-FPN for upsampling the image features extracted by SAM. We employ wavelet decomposition to denoise the depth-guided images for further noise reduction and weather interference. At last, we employ self-attention mechanisms to adaptively reweight the fused features, enhancing informative features while suppressing excess noise. In summary, RoboFusion significantly reduces noise by leveraging the generalization and robustness of VFMs, thereby enhancing the resilience of multi-modal 3D object detection. Consequently, RoboFusion achieves SOTA performance in noisy scenarios, as demonstrated by the KITTI-C and nuScenes-C benchmarks. Code is available at https://github.com/adept-thu/RoboFusion.

Xiaobo Shen, Qianxin Huang, Long Lan, Yuhui Zheng

As video-based social networks continue to grow exponentially, there is a rising interest in video retrieval using natural language. Cross-modal hashing, which learns compact hash code for encoding multi-modal data, has proven to be widely effective in large-scale cross-modal retrieval, e.g., image-text retrieval, primarily due to its computation and storage efficiency. However, when applied to video-text retrieval, existing cross-modal hashing methods generally extract features at the frame- or word-level for videos and texts individually, thereby ignoring their long-term dependencies. To address this issue, we propose Contrastive Transformer Cross-Modal Hashing (CTCH), a novel approach designed for video-text retrieval task. CTCH employs bidirectional transformer encoder to encode video and text and leverages their long-term dependencies. CTCH further introduces supervised multi-modality contrastive loss that effectively exploits inter-modality and intra-modality similarities among videos and texts. The experimental results on three video benchmark datasets demonstrate that CTCH outperforms the state-of-the-arts in video-text retrieval tasks.

Xiaobo Shen, Haoyu Cai, Xiuwen Gong, Yuhui Zheng

Hashing utilizes hash code as a compact image representation, offering excellent performance in large-scale image retrieval due to its computational and storage advantages. However, the prevalence of degraded images on social media platforms, resulting from imperfections in the image capture process, poses new challenges for conventional image retrieval methods. To address this issue, we propose Contrastive Transformer Masked Image Hashing (CTMIH), a novel deep unsupervised hashing method specifically designed for degraded image retrieval, which is challenging yet relatively less studied. CTMIH addresses the problem by training on transformed and masked images, aiming to learn transform-invariant hash code in an unsupervised manner to mitigate performance degradation caused by image deterioration. CTMIH utilizes Vision Transformer (ViT) architecture applied to image patches to capture distant semantic relevance. CTMIH introduces cross-view debiased contrastive loss to align hash tokens from augmented views of the same image and presents semantic mask reconstruction loss at the patch level to recover masked patch tokens. Extensive empirical studies conducted on three benchmark datasets demonstrate the superiority of the proposed CTMIH over the state-of-the-art in both degraded and normal image retrieval.

Yingnan Liu, Yingtian Zou, Rui Qiao, Fusheng Liu, Mong Li Lee, Wynne Hsu

Domain generalization aims to develop models that are robust to distribution shifts. Existing methods focus on learning invariance across domains to enhance model robustness, and data augmentation has been widely used to learn invariant predictors, with most methods performing augmentation in the input space. However, augmentation in the input space has limited diversity whereas in the feature space is more versatile and has shown promising results. Nonetheless, feature semantics is seldom considered and existing feature augmentation methods suffer from a limited variety of augmented features. We decompose features into class-generic, class-specific, domain-generic, and domain-specific components. We propose a cross-domain feature augmentation method named XDomainMix that enables us to increase sample diversity while emphasizing the learning of invariant representations to achieve domain generalization. Experiments on widely used benchmark datasets demonstrate that our proposed method is able to achieve state-of-the-art performance. Quantitative analysis indicates that our feature augmentation approach facilitates the learning of effective models that are invariant across different domains.

Jiafeng Liang, Shixin Jiang, Zekun Wang, Haojie Pan, Zerui Chen, Zheng Chu, Ming Liu, Ruiji Fu, Zhongyuan Wang, Bing Qin

There are substantial instructional videos on the Internet, which provide us tutorials for completing various tasks. Existing instructional video datasets only focus on specific steps at the video level, lacking experiential guidelines at the task level, which can lead to beginners struggling to learn new tasks due to the lack of relevant experience. Moreover, the specific steps without guidelines are trivial and unsystematic, making it difficult to provide a clear tutorial. To address these problems, we present the Guide (Guideline-Guided) dataset, which contains 3.5K videos of 560 instructional tasks in 8 domains related to our daily life. Specifically, we annotate each instructional task with a guideline, representing a common pattern shared by all task-related videos. On this basis, we annotate systematic specific steps, including their associated guideline steps, specific step descriptions and timestamps. Our proposed benchmark consists of three sub-tasks to evaluate comprehension ability of models: (1) Step Captioning: models have to generate captions for specific steps from videos. (2) Guideline Summarization: models have to mine the common pattern in task-related videos and summarize a guideline from them. (3) Guideline-Guided Captioning: models have to generate captions for specific steps under the guide of guideline. We evaluate plenty of foundation models with Guide and perform in-depth analysis. Given the diversity and practicality of Guide, we believe that it can be used as a better benchmark for instructional video comprehension.

Yimin Jiang, Huibing Wang, Jinjia Peng, Xianping Fu, Yang Wang

Person search aims to localize specific a target person from a gallery set of images with various scenes. As the scene of moving pedestrian changes, the captured person image inevitably bring in lots of background noise and foreground noise on the person feature, which are completely unrelated to the person identity, leading to severe performance degeneration. To address this issue, we present a Scene-Adaptive Person Search (SEAS) model by introducing bilateral modulations to simultaneously eliminate scene noise and maintain a consistent person representation to adapt to various scenes. In SEAS, a Background Modulation Network (BMN) is designed to encode the feature extracted from the detected bounding box into a multi-granularity embedding, which reduces the input of background noise from multiple levels with norm-aware. Additionally, to mitigate the effect of foreground noise on the person feature, SEAS introduces a Foreground Modulation Network (FMN) to compute the clutter reduction offset for the person embedding based on the feature map of the scene image. By bilateral modulations on both background and foreground within an end-to-end manner, SEAS obtains consistent feature representations without scene noise. SEAS can achieve state-of-the-art (SOTA) performance on two benchmark datasets, CUHK-SYSU with 97.1% mAP and PRW with 60.5% mAP. The code is available at https://github.com/whbdmu/SEAS.

Deyi Ji, Wenwei Jin, Hongtao Lu, Feng Zhao

The ascension of Unmanned Aerial Vehicles (UAVs) in various fields necessitates effective UAV image segmentation, which faces challenges due to the dynamic perspectives of UAV-captured images. Traditional segmentation algorithms falter as they cannot accurately mimic the complexity of UAV perspectives, and the cost of obtaining multi-perspective labeled datasets is prohibitive. To address these issues, we introduce the PPTFormer, a novel Pseudo Multi-Perspective Transformer network that revolutionizes UAV image segmentation. Our approach circumvents the need for actual multi-perspective data by creating pseudo perspectives for enhanced multi-perspective learning. The PPTFormer network boasts Perspective Decomposition, novel Perspective Prototypes, and a specialized encoder and decoder that together achieve superior segmentation results through Pseudo Multi-Perspective Attention (PMP Attention) and fusion. Our experiments demonstrate that PPTFormer achieves state-of-the-art performance across five UAV segmentation datasets, confirming its capability to effectively simulate UAV flight perspectives and significantly advance segmentation precision. This work presents a pioneering leap in UAV scene understanding and sets a new benchmark for future developments in semantic segmentation.

Haodong Hong, Sen Wang, Zi Huang, Qi Wu, Jiajun Liu

Current Vision-and-Language Navigation (VLN) tasks mainly employ textual instructions to guide agents. However, being inherently abstract, the same textual instruction can be associated with different visual signals, causing severe ambiguity and limiting the transfer of prior knowledge in the vision domain from the user to the agent. To fill this gap, we propose Vision-and-Language Navigation with Multi-modal Prompts (VLN-MP), a novel task augmenting traditional VLN by integrating both natural language and images in instructions. VLN-MP not only maintains backward compatibility by effectively handling text-only prompts but also consistently shows advantages with different quantities and relevance of visual prompts. Possible forms of visual prompts include both exact and similar object images, providing adaptability and versatility in diverse navigation scenarios. To evaluate VLN-MP under a unified framework, we implement a new benchmark that offers: (1) a training-free pipeline to transform textual instructions into multi-modal forms with landmark images; (2) diverse datasets with multi-modal instructions for different downstream tasks; (3) a novel module designed to process various image prompts for seamless integration with state-of-the-art VLN models. Extensive experiments on four VLN benchmarks (R2R, RxR, REVERIE, CVDN) show that incorporating visual prompts would significantly boost navigation performance. While maintaining efficiency with text-only prompts, VLN-MP enables agents to navigate in the pre-explore setting and outperform text-based models, showing its broader applicability. Code is available at https://github.com/honghd16/VLN-MP.