Adversarial attack is a major obstacle to the deployment of deep neural networks (DNNs) for security-sensitive applications. To address these adversarial perturbations, various adversarial defense strategies have been developed, with Adversarial Training (AT) being one of the most effective methods to protect neural networks from adversarial attacks. However, existing AT methods struggle against training-agnostic attacks due to their limited generalizability. This suggests that the AT models lack a unified perspective for various attacks to conduct universal defense. This paper sheds light on a generalizable prior under various attacks: consistent class confusion (3C), i.e., an AT classifier often confuses the predictions between correct and ambiguous classes in a highly similar pattern among diverse attacks. Relying on this latent prior as a bridge between seen and agnostic attacks, we propose a more generalized AT model by mitigating consistent class confusion (M3C) to resist training-agnostic attacks. Specifically, we optimize an Adversarial Confusion Loss (ACL), which is weighted by uncertainty, to distinguish the most confused classes and encourage the AT model to focus on these confused samples. To suppress malignant features affecting correct predictions and producing significant class confusion, we propose a Gradient-Aware Attention (GAA) mechanism to enhance the classification confidence of correct classes and eliminate class confusion. Experiments on multiple benchmarks and network frameworks demonstrate that our M3C model significantly improves the generalization of AT robustness against agnostic attacks. The finding of the 3C prior reveals the potential and possibility for defending against a wide range of attacks, and provides a new perspective to overcome such challenge in this field.

Unsupervised Domain Adaptation (UDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain by learning domain-invariant representations. Motivated by the recent success of Vision Transformers (ViTs), several UDA approaches have adopted ViT architectures to exploit fine-grained patch-level representations, which are unified as Transformer-based $D$ omain $A$ daptation (TransDA) independent of CNN-based. However, we have a key observation in TransDA: due to inherent domain shifts, patches (tokens) from different semantic categories across domains may exhibit abnormally high similarities, which can mislead the self-attention mechanism and degrade adaptation performance. To solve that, we propose a novel $P$ atch- $A$ daptation Transformer (PATrans), which first identifies similarity-anomalous patches and then adaptively suppresses their negative impact to domain alignment, i.e. token calibration. Specifically, we introduce a $P$ atch- $A$ daptation $A$ ttention (PAA) mechanism to replace the standard self-attention mechanism, which consists of a weight-shared triple-branch mixed attention mechanism and a patch-level domain discriminator. The mixed attention integrates self-attention and cross-attention to enhance intra-domain feature modeling and inter-domain similarity estimation. Meanwhile, the patch-level domain discriminator quantifies the anomaly probability of each patch, enabling dynamic reweighting to mitigate the impact of unreliable patch correspondences. Furthermore, we introduce a contrastive attention regularization strategy, which leverages category-level information in a contrastive learning framework to promote class-consistent attention distributions. Extensive experiments on four benchmark datasets demonstrate that PATrans attains significant improvements over existing state-of-the-art UDA methods (e.g., 89.2% on the VisDA-2017). Code is available at: https://github.com/YSY145/PATrans.

Existing defense methods fail to defend against un known attacks and thus raise generalization issue of adversarial robustness. To remedy this problem, we attempt to delve into some underlying common characteristics among various attacks for generality. In this work, we reveal the commonly overlooked low entropy prior (LE) implied in various adversarial samples, and shed light on the universal robustness against unseen attacks in inference phase. LE prior is elaborated as two properties across various attacks as shown in Fig. 1 and 2: 1) low entropy misclassification for adversarial samples and 2) lower entropy prediction for higher attack intensity. This phenomenon stands in stark contrast to the naturally distributed samples. The LE prior can instruct existing test-time defense methods, thus we propose a two-stage REAL approach: Rectify Adversarial sample based on LE prior for test-time adversarial rectification. Specifically, to align adversarial samples more closely with clean samples, we propose to first rectify adversarial samples misclassified with low entropy by reverse maximizing prediction entropy, thereby eliminating their adversarial nature. To ensure the rectified samples can be correctly classified with low entropy, we carry out secondary rectification by forward minimizing prediction entropy, thus creating a Max-Min entropy optimization scheme. Further, based on the second property, we propose an attack aware weighting mechanism to adaptively adjust the strengths of Max-Min entropy objectives. Experiments on several datasets show that REAL can greatly improve the performance of existing sample rectification models.

State space model (SSM) is a mathematical model used to describe and analyze the behavior of dynamic systems. This model has witnessed numerous applications in several fields, including control theory, signal processing, economics, and machine learning. In the field of deep learning, SSMs are used to process sequence data, such as time series analysis, natural language processing (NLP), and video understanding. By mapping sequence data to state space, long-term dependencies in the data can be better captured. In particular, modern SSMs have shown strong representational capabilities in NLP, especially in long sequence modeling, while maintaining linear time complexity. In particular, based on the latest SSMs, Mamba merges time-varying parameters into SSMs toward efficient training and inference. Given its impressive efficiency and strong long-range dependency modeling capability, Mamba is expected to become a new AI architecture that may be capable of surpassing Transformer. Recently, a number of works attempt to study the potential of Mamba in various fields, such as general vision, multimodal learning, medical image analysis, and remote sensing image analysis, by extending Mamba from natural language domain to visual domain. To fully understand Mamba in the visual domain, we conduct a comprehensive survey and present a taxonomy study. This survey focuses on Mamba's application to a variety of visual tasks and data types, and discusses its predecessors, recent advances, and far-reaching impact on a wide range of domains.

Mixed-Modal Image Retrieval (MMIR) as a flexible search paradigm has attracted wide attention. However, previous approaches always achieve limited performance, due to two critical factors are seriously overlooked. 1) The contribution of image and text modalities is different, but incorrectly treated equally. 2) There exist inherent labeling noises in describing users' intentions with text in web datasets from diverse real-world scenarios, giving rise to overfitting. We propose a Dynamic Weighted Combiner (DWC) to tackle the above challenges, which includes three merits. First, we propose an Editable Modality De-equalizer (EMD) by taking into account the contribution disparity between modalities, containing two modality feature editors and an adaptive weighted combiner. Second, to alleviate labeling noises and data bias, we propose a dynamic soft-similarity label generator (SSG) to implicitly improve noisy supervision. Finally, to bridge modality gaps and facilitate similarity learning, we propose a CLIP-based mutual enhancement module alternately trained by a mixed-modality contrastive loss. Extensive experiments verify that our proposed model significantly outperforms state-of-the-art methods on real-world datasets. The source code is available at https://github.com/fuxianghuang1/DWC.

Unsupervised domain adaptation (UDA) intends to transfer knowledge from a labeled source domain to an unlabeled target domain. Many current methods focus on learning feature representations that are both discriminative for classification and invariant across domains by simultaneously optimizing domain alignment and classification tasks. However, these methods often overlook a crucial challenge: the inherent conflict between these two tasks during gradient-based optimization. In this paper, we delve into this issue and introduce two effective solutions known as Gradient Harmonization, including GH and GH++, to mitigate the conflict between domain alignment and classification tasks. GH operates by altering the gradient angle between different tasks from an obtuse angle to an acute angle, thus resolving the conflict and trade-offing the two tasks in a coordinated manner. Yet, this would cause both tasks to deviate from their original optimization directions. We thus further propose an improved version, GH++, which adjusts the gradient angle between tasks from an obtuse angle to a vertical angle. This not only eliminates the conflict but also minimizes deviation from the original gradient directions. Finally, for optimization convenience and efficiency, we evolve the gradient harmonization strategies into a dynamically weighted loss function using an integral operator on the harmonized gradient. Notably, GH/GH++ are orthogonal to UDA and can be seamlessly integrated into most existing UDA models. Theoretical insights and experimental analyses demonstrate that the proposed approaches not only enhance popular UDA baselines but also improve recent state-of-the-art models.

Generalization of the person re-identification (ReID) model plays an important role in practical application, and we discuss a simple yet effective regularizer to improve it inspired by Adversarial Training (AT). AT has been indicated as an advanced regularizer due to its adversarial mechanism, ability to mine hard samples, and nature of data augmentation. However, serving as an augmentation-based regularizer, AT shows low diversity of the perturbation, excessive computational cost, and the optimization dilemma between adversarial robustness and accuracy for ReID task, and is thus suboptimal. To tackle these limitations and get a more effective regularizer for ReID, we rethink the nature of AT and unveil that the adversarial data augmentation is essentially reflected by gradients. Based on this, a novel implicit regularizer, named Stochastic Gradient Perturbation (SGP), is proposed, which naturally brings three merits: 1) Better diversity of the perturbation due to the proposed non-directional stochastic perturbations rather than directional adversarial perturbations. 2) Lower computational cost due to the proposed implicit gradient augmentation rather than explicitly additional data. 3) The optimization dilemma of the adversarial robustness and generalization is naturally overcome since SGP contains the adversarial gradient perturbation. Further, we put forward a perspective that the generalization and adversarial robustness may have an inter unity. Experiments on the baseline and SOTA models demonstrate powerful performances of the plugged-played SGP, and both generalization and adversarial robustness can be guaranteed.

Image Retrieval with Text Feedback (IRTF) is an emerging research topic where the query consists of an image and a text expressing a requested attribute modification. The goal is to retrieve the target images similar to the query text modified query image. The existing methods usually adopt feature fusion of the query image and text to match the target image. However, they ignore two crucial issues: overfitting and low diversity of training data, which make the feature fusion based IRTF task not generalizable. Conventional generation based data augmentation is an effective way to alleviate overfitting and improve diversity, but increases the volume of training data and generation model parameters, which is bound to bring huge computation costs. By rethinking the conventional data augmentation mechanism, we propose a plug-and-play Gradient Augmentation (GA) based regularization approach. Specifically, GA contains two items: 1) To alleviate model overfitting on the training set, we deduce anexplicit adversarial gradient augmentationfrom the perspective of adversarial training, which challenges the“no free lunch”philosophy. 2) To improve the diversity of training set, we propose animplicit isotropic gradient augmentationfrom the perspective of gradient descent-based optimization, which achieves the goal ofbig gain but no pain. Besides, we introduce deep metric learning to train the model and provide theoretical insights of GA on generalisation. Finally, we propose a new evaluation protocol called Weighted Harmonic Mean (WHM) to assess the model generalisation. Experiments show that our GA outperforms the state-of-the-art methods by 6.2 and 4.7% on CSS and Fashion200 k datasets, respectively, without bells and whistles.

RGB-Infrared cross-modal person re-identification (Re-ID) has drawn increasing attention due to its application value in practice. Most of the current works rely on a supervised training manner. However, in real-world applications, manual collection of pair-wise RGB-Infrared (IR) person data is labor-intensive and time-consuming. Moreover, when a trained model is directly used in another domain, there is usually a significant performance drop. To overcome the above problems, we make the first attempt to transfer the learned model to a new RGB-IR domain which is unlabeled. The practical problem covers two kinds of challenges, i.e., cross-modal (RGB-Infrared) and cross-domain (different dataset) person Re-ID. Previous works have often considered only one of them either cross-modal or cross-domain. In this work, we propose a dual alignment network (DAN) to solve the RGB-Infrared cross-modal cross-domain person Re-ID problem. This network consists of three parts: Domain Adversarial Alignment component (DAA), Pseudo Label Generation module for target domain (PLG), and Cross-Modal Alignment component (CMA). These three modules complement and promote the model to learn domain-invariant and modality-invariant person representations. Further, we propose a protocol of cross-modal cross-domain person Re-ID by synthesizing target domains by adding random noise, adjusting the lighting intensity, and changing the background color, respectively. Experiments on real and synthetic datasets under the same cross-modalities across domains demonstrate the effectiveness of our method.

Learning to hash has been widely applied for image retrieval due to the low storage and high retrieval efficiency. Existing hashing methods assume that the distributions of the retrieval pool (i.e., the data sets being retrieved) and the query data are similar, which, however, cannot truly reflect the real-world condition due to the unconstrained visual cues, such as illumination, pose, background, and so on. Due to the large distribution gap between the retrieval pool and the query set, the performances of traditional hashing methods are seriously degraded. Therefore, we first propose a new efficient but transferable hashing model for unconstrained cross-domain visual retrieval, in which the retrieval pool and the query sample are drawn from different but semantic relevant domains. Specifically, we propose a simple yet effective unsupervised hashing method, domain adaptation preconceived hashing (DAPH), toward learning domain-invariant hashing representation. Three merits of DAPH are observed: 1) to the best of our knowledge, we first propose unconstrained visual retrieval by introducing DA into hashing for learning transferable hashing codes; 2) a domain-invariant feature transformation with marginal discrepancy distance minimization and feature reconstruction constraint is learned, such that the hashing code is not only domain adaptive but content preserved; and 3) a DA preconceived quantization loss is proposed, which further guarantees the discrimination of the learned hashing code for sample retrieval. Extensive experiments on various benchmark data sets verify that our DAPH outperforms many state-of-the-art hashing methods toward unconstrained (unrestricted) instance retrieval in both single- and cross-domain scenarios.

Domain adaptive image retrieval includes single-domain retrieval and cross-domain retrieval. Most of the existing image retrieval methods only focus on single-domain retrieval, which assumes that the distributions of retrieval databases and queries are similar. However, in practical application, the discrepancies between retrieval databases often taken in ideal illumination/pose/background/camera conditions and queries usually obtained in uncontrolled conditions are very large. In this paper, considering the practical application, we focus on challenging cross-domain retrieval. To address the problem, we propose an effective method named Probability Weighted Compact Feature Learning (PWCF), which provides inter-domain correlation guidance to promote cross-domain retrieval accuracy and learns a series of compact binary codes to improve the retrieval speed. First, we derive our loss function through the Maximum A Posteriori Estimation (MAP): Bayesian Perspective (BP) induced focal-triplet loss, BP induced quantization loss and BP induced classification loss. Second, we propose a common manifold structure between domains to explore the potential correlation across domains. Considering the original feature representation is biased due to the inter-domain discrepancy, the manifold structure is difficult to be constructed. Therefore, we propose a new feature named Histogram Feature of Neighbors (HFON) from the sample statistics perspective. Extensive experiments on various benchmark databases validate that our method outperforms many state-of-the-art image retrieval methods for domain adaptive image retrieval. The source code is available at {https://github.com/fuxianghuang1/PWCF}.

Recently, learning to hash has been widely studied for image retrieval thanks to the computation and storage efficiency of binary codes. Most existing learning to hash methods have yielded significant performance. However, for most existing learning to hash methods, sufficient training images are required and used to learn precise hashing codes. In some real-world applications, there are not always sufficient training images in the domain of interest. In addition, some existing supervised approaches need a amount of labeled data, which is an expensive process in terms of time, labor and human expertise. To handle such problems, inspired by transfer learning, we propose a simple yet effective unsupervised hashing method named Optimal Projection Guided Transfer Hashing (GTH) where we borrow the images of other different but related domain i.e., source domain to help learn precise hashing codes for the domain of interest i.e., target domain. In GTH, we aim to learn domain-invariant hashing functions. To achieve that, we propose to minimize the error matrix between two hashing projections of target and source domains. We seek for the maximum likelihood estimation (MLE) solution of the error matrix between the two hashing projections due to the domain gap. Furthermore, an alternating optimization method is adopted to obtain the two projections of target and source domains. By doing so, two projections can be progressively aligned. Extensive experiments on various benchmark databases for cross-domain visual recognition verify that our method outperforms many state-of-the-art learning to hash methods. The source code is available at https://github.com/liuji93/GTH.

Existing hashing methods have yielded significant performance in image and multimedia retrieval, which can be categorized into two groups: shallow hashing and deep hashing. However, there still exist some intrinsic limitations among them. The former generally adopts a one-step strategy to learn the hashing codes for discovering the discriminative binary feature, but the latent discriminative information in the learned hashing codes is not well exploited. The latter, as deep neural network based hashing models, can learn highly discriminative and compact features, but relies on large-scale data and computation resources for numerous network parameters tuning with back-propagation optimization. Straightforward training of deep hashing models from scratch on small-scale data is almost impossible. Therefore, in order to develop efficient but effective learning to hash algorithm that depends only on small-scale data, we propose a novel non-neural network based deep-like learning framework, i.e. multi-level cascaded hashing (MCH) approach with hierarchical learning strategy, for image retrieval. The contributions are threefold. First, a hashing-in-hash architecture is designed in MCH, which inherits the excellent traits of traditional neural networks based deep learning, such that discriminative binary features that are beneficial to image retrieval can be effectively captured. Second, in each level the binary features of all preceding levels and the visual appearance feature are simultaneously cascaded as inputs of all subsequent levels to retrain, which fully exploits the implicated discriminative information. Third, a basic learning to hash (BLH) model with label constraint is proposed for hierarchical learning. Without loss of generality, the existing hashing models can be easily integrated into our MCH framework. We show experimentally on small- and large-scale visual retrieval tasks that our method outperforms several state-of-the-arts.

In this study, the daily AIRS total ozone and temperature profile in 2015 are used to investigate the spatiotemporal variation characteristics of the correlations between the Arctic atmospheric ozone and temperature in the lower stratosphere and tropopshere. The results show that a remarkable seasonal split exists in the correlation between the Arctic ozone and temperature. In the lower stratosphere, the ozone and temperature present a positive correlation in winter and summer, while showing weak or not significant correlation in spring and autumn. In the troposphere, the ozone and temperature have a strong negative correlation in spring and autumn, while presenting a weak or not significant negative correlation in winter and summer. In the stratosphere, the biggest positive correlation coefficients are 0.8 and above at the Barents Sea in summer, and 0.6 to 0.8 at the Baffin Bay in winter. In troposphere, the biggest negative correlation coefficients reach -0.9 and above at the Greenland-Queen Elizabeth Islands in spring. As for the vertical distribution of the correlation between the Arctic ozone and temperature, the correlation coefficients are positive in lower stratosphere, negative in troposphere and zero at tropopause and the distribution varies with seasons and regions dramatically.

The Solar Backscatter Ultraviolet Sounder (SBUS) is one of the 11 main payload instruments onboard Feng Yun-3 (FY-3), the second generation of Chinese polar orbit meteorological satellites. This paper presents the results of SBUS instrument calibration, and data and product validation during the prelaunch and postlaunch periods. Topics include the instrument prelaunch calibration and characterization, in-orbit monitoring, validation of the ozone profiles retrieved from the FY-3 SBUS measurements, and an application of the retrievals to monitoring the 2011 Arctic ozone depletion. For the prelaunch calibration of SBUS, the estimated uncertainty of laboratory calibration is approximately 4.7%. The in-orbit solar irradiance measurements indicate that the diffuser reflectivity degraded approximately 15% for the 252-nm channel, and 3% to 5% for the other 11 channels during a 12-mo period. Using ozone vertical profiles retrieved from National Oceanic and Atmospheric Administration Solar Backscatter Ultraviolet (SBUV)/2s as a “truth,” the initial comparison of ozone profiles between FY-3 SBUS and SBUV/2s finds that the relative percent bias of the FY-3 SBUS with the SBUV/2 results is good. The averaged differences range over to ±7% for FY-3A SBUS and ±6% for FY-3B SBUS. The SBUS ozone profile retrievals reveal that the spring 2011 Arctic ozone depletion mainly resulted from a sharp ozone decrease in the upper troposphere to lower stratosphere, which accounts for 70% to 80% of the total ozone loss.

A cross-sensor calibration technique is developed and applied to improve upon the prelaunch radiance calibration and characterization for the Total Ozone Unit (TOU) onboard the FengYun-3/A satellite. The Level 3 products from the National Aeronautics and Space Administration Ozone Monitoring Instrument (OMI) onboard the Earth Observing System Aura are used as input to a radiative transfer model to predict the TOU radiances and characterize the biases for the measurements over the Pacific Ocean in low- and midlatitudes. The coefficients are derived from a regression algorithm to adjust the TOU radiances. It is shown that, after the measurement bias correction, the biases between the retrieved total column ozone products from the TOU with those from the OMI Total Ozone Mapping Spectrometer (TOMS)-Version 8 products and those from a set of ground-based station measurements are 3 % and 5% , respectively. The variations in the estimated total ozone amounts from the TOU are consistent with those derived from Solar Backscatter Ultraviolet Radiometer instruments and OMI for a period from January 2010 to February 2011.