Unlabeled Target Domain Research Articles

Unsupervised Domain Adaptation by Multi-Loss Gap Minimization Learning for Person Re-Identification

Unsupervised domain adaptation (UDA) person re-identification (ReID) faces enormous challenges due to the severe shift between the source and target domains, as well as the dramatic variations within the target domain. In this paper, to address these issues, we propose a multi-loss gap minimization learning (MGML) approach for UDA person ReID. Firstly, we introduce the part model to learn discriminative patch features and design a Patch-based Part Ignoring (PPI) loss to select reliable instances for the efficient learning of the part model. Then, given the gap that typically occurs because of the inter-domain shift and intra-domain variations, a Gap-based Minimum Camera Discrepancy (G-MCD) loss is proposed. Specifically, in terms of the inter-domain, we propose to leverage the tracklet and camera information to label each distance vector, and accordingly align pair-wise distance distributions to bridge the inter-domain gap. As for the intra-domain, to alleviate the biased search, we propose to perform the mined neighborhood for intra-camera and inter-camera separately to optimize matching results by exploring neighborhood relations more deeply. Finally, experimental results on three challenging datasets demonstrate that applying our method to unlabeled target domain outperforms current UDA methods for person ReID.

SECRET: Self-Consistent Pseudo Label Refinement for Unsupervised Domain Adaptive Person Re-identification

Unsupervised domain adaptive person re-identification aims at learning on an unlabeled target domain with only labeled data in source domain. Currently, the state-of-the-arts usually solve this problem by pseudo-label-based clustering and fine-tuning in target domain. However, the reason behind the noises of pseudo labels is not sufficiently explored, especially for the popular multi-branch models. We argue that the consistency between different feature spaces is the key to the pseudo labels’ quality. Then a SElf-Consistent pseudo label RefinEmenT method, termed as SECRET, is proposed to improve consistency by mutually refining the pseudo labels generated from different feature spaces. The proposed SECRET gradually encourages the improvement of pseudo labels’ quality during training process, which further leads to better cross-domain Re-ID performance. Extensive experiments on benchmark datasets show the superiority of our method. Specifically, our method outperforms the state-of-the-arts by 6.3% in terms of mAP on the challenging dataset MSMT17. In the purely unsupervised setting, our method also surpasses existing works by a large margin. Code is available at https://github.com/LunarShen/SECRET.

Reliability Exploration with Self-Ensemble Learning for Domain Adaptive Person Re-identification

Person re-identifcation (Re-ID) based on unsupervised domain adaptation (UDA) aims to transfer the pre-trained model from one labeled source domain to an unlabeled target domain. Existing methods tackle this problem by using clustering methods to generate pseudo labels. However, pseudo labels produced by these techniques may be unstable and noisy, substantially deteriorating models’ performance. In this paper, we propose a Reliability Exploration with Self-ensemble Learning (RESL) framework for domain adaptive person ReID. First, to increase the feature diversity, multiple branches are presented to extract features from different data augmentations. Taking the temporally average model as a mean teacher model, online label refning is conducted by using its dynamic ensemble predictions from different branches as soft labels. Second, to combat the adverse effects of unreliable samples in clusters, sample reliability is estimated by evaluating the consistency of different clusters’ results, followed by selecting reliable instances for training and re-weighting sample contribution within Re-ID losses. A contrastive loss is also utilized with cluster-level memory features which are updated by the mean feature. The experiments demonstrate that our method can signifcantly surpass the state-of-the-art performance on the unsupervised domain adaptive person ReID.

Denoised Maximum Classifier Discrepancy for Source-Free Unsupervised Domain Adaptation

Source-Free Unsupervised Domain Adaptation(SFUDA) aims to adapt a pre-trained source model to an unlabeled target domain without access to the original labeled source domain samples. Many existing SFUDA approaches apply the self-training strategy, which involves iteratively selecting confidently predicted target samples as pseudo-labeled samples used to train the model to fit the target domain. However, the self-training strategy may also suffer from sample selection bias and be impacted by the label noise of the pseudo-labeled samples. In this work, we provide a rigorous theoretical analysis on how these two issues affect the model generalization ability when applying the self-training strategy for the SFUDA problem. Based on this theoretical analysis, we then propose a new Denoised Maximum Classifier Discrepancy (D-MCD) method for SFUDA to effectively address these two issues. In particular, we first minimize the distribution mismatch between the selected pseudo-labeled samples and the remaining target domain samples to alleviate the sample selection bias. Moreover, we design a strong-weak self-training paradigm to denoise the selected pseudo-labeled samples, where the strong network is used to select pseudo-labeled samples while the weak network helps the strong network to filter out hard samples to avoid incorrect labels. In this way, we are able to ensure both the quality of the pseudo-labels and the generalization ability of the trained model on the target domain. We achieve state-of-the-art results on three domain adaptation benchmark datasets, which clearly validates the effectiveness of our proposed approach. Full code is available at https://github.com/kkkkkkon/D-MCD.

Mutual Nearest Neighbor Contrast and Hybrid Prototype Self-Training for Universal Domain Adaptation

Universal domain adaptation (UniDA) aims to transfer knowledge learned from a labeled source domain to an unlabeled target domain under domain shift and category shift. Without prior category overlap information, it is challenging to simultaneously align the common categories between two domains and separate their respective private categories. Additionally, previous studies utilize the source classifier's prediction to obtain various known labels and one generic "unknown" label of target samples. However, over-reliance on learned classifier knowledge is inevitably biased to source data, ignoring the intrinsic structure of target domain. Therefore, in this paper, we propose a novel two-stage UniDA framework called MATHS based on the principle of mutual nearest neighbor contrast and hybrid prototype discrimination. In the first stage, we design an efficient mutual nearest neighbor contrastive learning scheme to achieve feature alignment, which exploits the instance-level affinity relationship to uncover the intrinsic structure of two domains. We introduce a bimodality hypothesis for the maximum discriminative probability distribution to detect the possible target private samples, and present a data-based statistical approach to separate the common and private categories. In the second stage, to obtain more reliable label predictions, we propose an incremental pseudo-classifier for target data only, which is driven by the hybrid representative prototypes. A confidence-guided prototype contrastive loss is designed to optimize the category allocation uncertainty via a self-training mechanism. Extensive experiments on three benchmarks demonstrate that MATHS outperforms previous state-of-the-arts on most UniDA settings.

Evidential Neighborhood Contrastive Learning for Universal Domain Adaptation

Universal domain adaptation (UniDA) aims to transfer the knowledge learned from a labeled source domain to an unlabeled target domain without any constraints on the label sets. However, domain shift and category shift make UniDA extremely challenging, mainly attributed to the requirement of identifying both shared “known” samples and private “unknown” samples. Previous methods barely exploit the intrinsic manifold structure relationship between two domains for feature alignment, and they rely on the softmax-based scores with class competition nature to detect underlying “unknown” samples. Therefore, in this paper, we propose a novel evidential neighborhood contrastive learning framework called TNT to address these issues. Specifically, TNT first proposes a new domain alignment principle: semantically consistent samples should be geometrically adjacent to each other, whether within or across domains. From this criterion, a cross-domain multi-sample contrastive loss based on mutual nearest neighbors is designed to achieve common category matching and private category separation. Second, toward accurate “unknown” sample detection, TNT introduces a class competition-free uncertainty score from the perspective of evidential deep learning. Instead of setting a single threshold, TNT learns a category-aware heterogeneous threshold vector to reject diverse “unknown” samples. Extensive experiments on three benchmarks demonstrate that TNT significantly outperforms previous state-of-the-art UniDA methods.

GearNet: Stepwise Dual Learning for Weakly Supervised Domain Adaptation

This paper studies a weakly supervised domain adaptation (WSDA) problem, where we only have access to the source domain with noisy labels, from which we need to transfer useful information to the unlabeled target domain. Although there have been a few studies on this problem, most of them only exploit unidirectional relationships from the source domain to the target domain. In this paper, we propose a universal paradigm called GearNet to exploit bilateral relationships between the two domains. Specifically, we take the two domains as different inputs to train two models alternately, and a symmetrical Kullback-Leibler loss is used for selectively matching the predictions of the two models in the same domain. This interactive learning schema enables implicit label noise canceling and exploit correlations between the source and target domains. Therefore, our GearNet has the great potential to boost the performance of a wide range of existing WSDA methods. Comprehensive experimental results show that the performance of existing methods can be significantly improved by equipping with our GearNet.

A multi camera unsupervised domain adaptation pipeline for object detection in cultural sites through adversarial learning and self-training

Object detection algorithms allow to enable many interesting applications which can be implemented in different devices, such as smartphones and wearable devices. In the context of a cultural site, implementing these algorithms in a wearable device, such as a pair of smart glasses, allow to enable the use of augmented reality (AR) to show extra information about the artworks and enrich the visitors’ experience during their tour. However, object detection algorithms require to be trained on many well annotated examples to achieve reasonable results. This brings a major limitation since the annotation process requires human supervision which makes it expensive in terms of time and costs. A possible solution to reduce these costs consist in exploiting tools to automatically generate synthetic labeled images from a 3D model of the site. However, models trained with synthetic data do not generalize on real images acquired in the target scenario in which they are supposed to be used. Furthermore, object detectors should be able to work with different wearable devices or different mobile devices, which makes generalization even harder. In this paper, we present a new dataset collected in a cultural site to study the problem of domain adaptation for object detection in the presence of multiple unlabeled target domains corresponding to different cameras and a labeled source domain obtained considering synthetic images for training purposes. We present a new domain adaptation method which outperforms current state-of-the-art approaches combining the benefits of aligning the domains at the feature and pixel level with a self-training process. We release the dataset at the following link https://iplab.dmi.unict.it/OBJ-MDA/ and the code of the proposed architecture at https://github.com/fpv-iplab/STMDA-RetinaNet.

Constraining pseudo‐label in self‐training unsupervised domain adaptation with energy‐based model

Deep learning is usually data starved, and the unsupervised domain adaptation (UDA) is developed to introduce the knowledge in the labeled source domain to the unlabeled target domain. Recently, deep self-training presents a powerful means for UDA, involving an iterative process of predicting the target domain and then taking the confident predictions as hard pseudo-labels for retraining. However, the pseudo-labels are usually unreliable, thus easily leading to deviated solutions with propagated errors. In this paper, we resort to the energy-based model and constrain the training of the unlabeled target sample with an energy function minimization objective. It can be achieved via a simple additional regularization or an energy-based loss. This framework allows us to gain the benefits of the energy-based model, while retaining strong discriminative performance following a plug-and-play fashion. The convergence property and its connection with classification expectation minimization are investigated. We deliver extensive experiments on the most popular and large-scale UDA benchmarks of image classification as well as semantic segmentation to demonstrate its generality and effectiveness.

Inter-cluster and intra-cluster joint optimization for unsupervised cross-domain person re-identification

The purpose of unsupervised cross-domain (UCD) person re-identification (re-ID) is to adapt the well pre-trained model on the labeled source domain to the unlabeled target domain, which tackles a more realistic problem. Most successful UCD methods are devoted to exploring the relationship at instance level in the target domain, but neglect the relationship at cluster level. However, person re-ID is the task of retrieving samples belonging to the same identity. Thus, exploring the cluster level relationship is critical to improve the performance of re-ID. In this paper, a novel inter-cluster and intra-cluster joint optimization (IcJO) model is proposed. Specifically, for the inter-cluster relationship, a novel centroid-based contrastive learning strategy is designed, making the sample close to its cluster centroid and far away from the heterogeneous centroids. In order to alleviate the influence of clustering noise, a cluster-based uncertainty estimation coefficient is introduced to guide the contrastive learning process. As for the intra-cluster relationship, a centroid-based maximum mean discrepancy (cMMD) loss is adopted to narrow the intra-cluster distribution. Moreover, in order to enhance the feature invariance of each identity, a centroid-based similarity loss is developed. Extensive experiments are conducted on three standard benchmarks. The results demonstrate that our IcJO model performs favorably against the state-of-the-art UCD person re-ID methods.

Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation.

Unsupervised domain adaptation (UDA) is an emerging technique that enables the transfer of domain knowledge learned from a labeled source domain to unlabeled target domains, providing a way of coping with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of privacy concerns about potential leaks in sensitive information contained in patient data, it is often challenging to share the data and labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on a few datasets and deep learning backbones, demonstrating the potential for our framework to be applied in challenging yet realistic clinical settings.

Deep Learning Domain Adaptation for Electro-Mechanical Actuator Fault Diagnosis Under Variable Driving Waveforms

Benefited from the sampling signal from monitoring sensors installed in electro-mechanical actuator (EMA), data-driven methods, such as deep learning models that can mine adaptive features from monitoring signal, have received wide attention in EMA fault diagnosis. A generalized hypothesis accepted by previous efforts on EMA fault diagnosis, is that target domain datasets (i.e., testing monitoring signal) are drawn from the same distribution with source domain datasets (i.e., training monitoring signal). However, except for the different working loads and velocities, various driving signal waveforms also exits in EMA, which further increase the domain discrepancy challenge and decrease the performance of those deep learning models. Therefore, this paper proposes a deep learning domain adaptation method that introduces joint maximum mean discrepancy (JMMD) into convolutional neural network (CNN), which aims to reduce the domain difference between labeled source domain datasets and unlabeled target domain datasets. In order to simulate domain discrepancy phenomenon, the NASA EMA datasets, sampled from acceleration sensors under variable driving waveforms with different working loads and velocities, are considered in this paper. Experiment results indicate that CNN-JMMD method can reduce the distribution difference and improve the diagnosis accuracy in source and target domains. In comparison with several available methods, extensive experimental evaluations on six domain tasks indicate the effectiveness of the proposed method. In addition, better performance has been obtained by selecting the proper trade-off factor between CNN and JMMD.

Transfer learning network for nuclear power plant fault diagnosis with unlabeled data under varying operating conditions

The performance of data-driven nuclear power plant fault diagnosis models trained by limited data cannot be guaranteed, the distribution discrepancy between training set (source domain) and test set (target domain) caused by varying operating conditions will seriously restrict their practical applications. To generalize the diagnostic knowledge learnt from labeled source domain to unlabeled target domain, a novel transfer learning method based on Maximum Mean Discrepancy (MMD) and Convolutional Neural Network (CNN) is proposed, which can reduce the domain discrepancy of the extracted features between source domain and target domain, by appending the MMD-based distribution discrepancy to the objective function of CNN. Numerical experiments with high-dimensional and strong-nonlinear complex nuclear power plant simulation data are conducted, results show that significant improvement in the diagnostic accuracy of target domain can be achieved on most transfer tasks. Besides, the influence of adopting different kernel functions (Linear kernel, Sigmoid kernel, Laplace kernel and Gaussian kernel) to calculate MMD is also studied, better transfer effect can be achieved when Gaussian kernel is used, including higher accuracy, faster convergence rate and less negative transfer. Overall, the proposed method is promising to expand the application scope of nuclear power plant intelligent fault diagnosis to varying operating conditions.

Dynamic Re-Weighting and Cross-Camera Learning for Unsupervised Person Re-Identification

Person Re-Identification (ReID) has witnessed tremendous improvements with the help of deep convolutional neural networks (CNN). Nevertheless, because different fields have their characteristics, most existing methods encounter the problem of poor generalization ability to invisible people. To address this problem, based on the relationship between the temporal and camera position, we propose a robust and effective training strategy named temporal smoothing dynamic re-weighting and cross-camera learning (TSDRC). It uses robust and effective algorithms to transfer valuable knowledge of existing labeled source domains to unlabeled target domains. In the target domain training stage, TSDRC iteratively clusters the samples into several centers and dynamically re-weights unlabeled samples from each center with a temporal smoothing score. Then, cross-camera triplet loss is proposed to fine-tune the source domain model. Particularly, to improve the discernibility of CNN models in the source domain, generally shared person attributes and margin-based softmax loss are adapted to train the source model. In terms of the unlabeled target domain, the samples are clustered into several centers iteratively and the unlabeled samples are dynamically re-weighted from each center. Then, cross-camera triplet loss is proposed to fine-tune the source domain model. Comprehensive experiments on the Market-1501 and DukeMTMC-reID datasets demonstrate that the proposed method vastly improves the performance of unsupervised domain adaptability.

Wasserstein Uncertainty Estimation for Adversarial Domain Matching.

Domain adaptation aims at reducing the domain shift between a labeled source domain and an unlabeled target domain, so that the source model can be generalized to target domains without fine tuning. In this paper, we propose to evaluate the cross-domain transferability between source and target samples by domain prediction uncertainty, which is quantified via Wasserstein gradient flows. Further, we exploit it for reweighting the training samples to alleviate the issue of domain shift. The proposed mechanism provides a meaningful curriculum for cross-domain transfer and adaptively rules out samples that contain too much domain specific information during domain adaptation. Experiments on several benchmark datasets demonstrate that our reweighting mechanism can achieve improved results in both balanced and partial domain adaptation.

ATPL: Mutually enhanced adversarial training and pseudo labeling for unsupervised domain adaptation

Unsupervised domain adaptation aims to transfer knowledge from a labeled source domain to a related but unlabeled target domain. Most existing approaches either adversarially reduce the domain shift or use pseudo-labels to provide category information during adaptation. However, an adversarial training method may sacrifice the discriminability of the target data, since no category information is available. Moreover, a pseudo labeling method is difficult to produce high-confidence samples, since the classifier is often source-trained and there exists the domain discrepancy. Thus, it may have a negative influence on learning target representations. A potential solution is to make them compensate each other to simultaneously guarantee the feature transferability and discriminability, which are the two key criteria of feature representations in domain adaptation. In this paper, we propose a novel method named ATPL, which mutually promotes Adversarial Training and Pseudo Labeling for unsupervised domain adaptation. ATPL can produce high-confidence pseudo-labels by adversarial training. Accordingly, ATPL will use the pseudo-labeled information to improve the adversarial training process, which can guarantee the feature transferability by generating adversarial data to fill in the domain gap. Those pseudo-labels can also boost the feature discriminability. Extensive experiments on real datasets demonstrate that the proposed ATPL method outperforms state-of-the-art unsupervised domain adaptation methods.

Multi-Source Selection Transfer Learning with Privacy-Preserving

Transfer learning has ability to create learning task of weakly labeled or unlabeled target domain by using knowledge of source domain to help, which can effectively improve the performance of target learning task. At present, the increased awareness of privacy protection restricts access to data sources and poses new challenges to the development of transfer learning. However, the research on privacy protection in transfer learning is very rare. The existing work mainly uses differential privacy technology and does not consider the distribution difference between data sources, or does not consider the conditional probability distribution of data, which causes negative transfer to harm the effect of algorithm. Therefore, this paper proposes multi-source selection transfer learning algorithm with privacy-preserving MultiSTLP, which is used in scenarios where target domain contains unlabeled data sets with only a small amount of group probability information and multiple source domains with a large number of labeled data sets. Group probability means that the class label of each sample in target data set is unknown, but the probability of each class in a given data group is available, and multiple source domains indicate that there are more than two source domains. The number of data set contains more than two data sets of source domain and one data set of target domain. The algorithm adapts to the marginal probability distribution and conditional probability distribution differences between domains, and can protect the privacy of target data and improve classification accuracy by fusing the idea of multi-source transfer learning and group probability into support vector machine. At the same time, it can select the representative dataset in source domains to improve efficiency relied on speeding up the training process of algorithm. Experimental results on several real datasets show the effectiveness of MultiSTLP, and it also has some advantages compared with the state-of-the-art transfer learning algorithm.

Attention-based cross-layer domain alignment for unsupervised domain adaptation

Unsupervised domain adaptation (UDA) aims to learn transferable knowledge from a labeled source domain and adapts a trained model to an unlabeled target domain. To bridge the gap between source and target domains, one prevailing strategy is to minimize the distribution discrepancy by aligning their semantic features extracted by deep models. The existing alignment-based methods mainly focus on reducing domain divergence in the same model layer. However, the same level of semantic information could distribute across model layers due to the domain shifts. To further boost model adaptation performance, we propose a novel method called Attention-based Cross-layer Domain Alignment (ACDA), which captures the semantic relationship between the source and target domains across model layers and calibrates each level of semantic information automatically through a dynamic attention mechanism. An elaborate attention mechanism is designed to reweight each cross-layer pair based on their semantic similarity for precise domain alignment, effectively matching each level of semantic information during model adaptation. Extensive experiments on multiple benchmark datasets consistently show that the proposed method ACDA yields state-of-the-art performance.

Unsupervised adversarial domain adaptation based on interpolation image for fish detection in aquaculture

Deep-learning-based object detection has brought great convenience to modern intelligent aquaculture. However, its good performance mainly depends on a huge number of labeled training data and the assumption that samples are independent and identically distributed. But in real intensive aquaculture, the acquisition of labeled samples in detection tasks is very time-consuming and laborious. At the same time, due to the influence of light, density, and background complexity, the domain shift caused by different data distribution is inevitable. To solve these problems, we introduce the domain adaptive object detection into aquaculture field for the first time to improve the cross-domain robustness of fish detection. To this end, the interpolation-based unsupervised adversarial domain adaptive fish detection model is proposed in this paper. The proposed model combines the detection network Faster RCNN and three adaptive modules to realize the cross-domain detection of fish. First, the style generation network is used to generate interpolation samples to fill the large distribution gap between the two aquaculture scenes. Then the alignment of local and global features is realized through different levels of adversarial training. Moreover, the focal loss is applied to the global adversarial loss to achieve better global feature alignment. This strategy increases the weight of hard-to-classify samples and enables the feature extraction network to extract more domain-invariant features. We conducted cross-domain adaptation experiments by collecting images from different aquaculture scenes. Compared with the original Faster RCNN and domain adaptation model DA-Faster, the proposed method can not only save the cost of manual annotation, but also effectively improve the detection performance of unlabeled target domain by leveraging the knowledge of the source domain.

Median Stable Clustering and Global Distance Classification for Cross-Domain Person Re-Identification

The person re-identification (ReID) method in a single-domain achieves appealing performance, but its reliance on label information greatly limits its extensibility. Therefore, the unsupervised cross-domain ReID method has received extensive attention. Its purpose is to optimize the model by using the labelled source domain and the unlabelled target domain and finally make the model well generalized in the target domain. We propose an unsupervised cross-domain ReID method based on median stable clustering (MSC) and global distance classification (GDC). Specifically, the measurement method used by MSC comprehensively considers the similarity between clusters, the number of samples in a cluster, and the combined similarity within a cluster. Different from the method based on triple loss, GDC can separate the distance distribution of positive and negative sample pairs in a global scope. In addition, considering that model performance is very sensitive to probability parameters when source domain memory is reconsolidated, we designed a dynamic memory reconsolidation (DMR) method to reduce the influence of parameters on performance. Extensive experiments on large-scale datasets (Market-1501, DukeMTMC-reID and MSMT17) demonstrate the superior performance of MSC-GDC over the state-of-the-art methods.