Unlabeled Data In Target Domain Research Articles

Distributionally robust unsupervised domain adaptation

Obtaining ground-truth label information from real-world data along with uncertainty quantification can be difficult or even infeasible. In the absence of labeled data for a certain task, unsupervised domain adaptation (UDA) techniques have shown great accomplishment by learning transferable knowledge from labeled source domain data and applying it to unlabeled target domain data, yet scarce studies consider addressing uncertainties under domain shifts to improve the model robustness. Distributionally robust learning (DRL) is emerging as a high-potential technique for building reliable learning systems that are robust to distribution shifts. In this study, we propose a distributionally robust unsupervised domain adaptation (DRUDA) method to enhance the machine learning model generalization ability under input space perturbations. The DRL-based UDA learning scheme is formulated as a min–max optimization problem by optimizing worst-case perturbations of the training source data. Our Wasserstein distributionally robust framework can reduce the shifts in the joint distributions across domains. The proposed DRUDA has been tested on digit datasets and the Office-31 dataset, compared with other state-of-the-art domain adaptation techniques. Our experimental results show that the proposed DRUDA leads to improvements in domain adaptation accuracy performance on target domains.

Cross-Modal Learning for Domain Adaptation in 3D Semantic Segmentation.

Domain adaptation is an important task to enable learning when labels are scarce. While most works focus only on the image modality, there are many important multi-modal datasets. In order to leverage multi-modality for domain adaptation, we propose cross-modal learning, where we enforce consistency between the predictions of two modalities via mutual mimicking. We constrain our network to make correct predictions on labeled data and consistent predictions across modalities on unlabeled target-domain data. Experiments in unsupervised and semi-supervised domain adaptation settings prove the effectiveness of this novel domain adaptation strategy. Specifically, we evaluate on the task of 3D semantic segmentation from either the 2D image, the 3D point cloud or from both. We leverage recent driving datasets to produce a wide variety of domain adaptation scenarios including changes in scene layout, lighting, sensor setup and weather, as well as the synthetic-to-real setup. Our method significantly improves over previous uni-modal adaptation baselines on all adaption scenarios. Code will be made available upon publication.

Modulation Classifier: A Few-Shot Learning Semi-Supervised Method Based on Multimodal Information and Domain Adversarial Network

This letter addresses the issue of underfitting or failure of deep learning models caused by insufficient training samples. Unlike previous supervised methods, a new few-shot learning semi-supervised automatic modulation recognition method based on multimodal information and domain adversarial network is proposed herein. The fusion input of multimodal information realizes the joint utilization of modulated signal modal features in the time and frequency domains. Domain adversarial training mines the potential knowledge information of a large number of unlabeled target domain data and introduces the convolutional block attention module (CBAM) to enhance the ability of the network to represent the key features of data. Numerical results validate the high-average classification accuracy of the proposed scheme compared to that of state-of-the-art schemes using fewer samples, particularly in high-order modulation classification. Alternative network structures are compared to confirm the applicability of multimodal information, domain adversarial training, and CBAM.

A Self-Supervised-Driven Open-Set Unsupervised Domain Adaptation Method for Optical Remote Sensing Image Scene Classification and Retrieval

Unsupervised domain adaptation (UDA) is an important solution to reduce the bias between the labeled source domain and the unlabeled target domain. It has attracted more attention for optical remote sensing image scene classification and retrieval. Currently, most of the previous work is devoted to closed-set UDA. In fact, the target domain often contains unknown classes. Moreover, some open UDA methods mine structural information of the target domain directly from the type knowledge of the source domain, and less directly from the unlabeled data of the target domain. In this paper, we propose a new self-supervised-driven open-set UDA method combining contrastive self-supervised learning with consistency self-training for optical remote sensing scene classification and retrieval. Specifically, a contrastive self-supervised learning network is introduced to learn discriminative features from the unlabeled target domain data. Moreover, a novel open-set class learning module is developed based on two-level confidence rules and the consistency self-training strategy, which can obtain reliable unknown class samples for co-training. Finally, an open-set dataset including six cross-domain scenarios is constructed based on three public datasets and several experiments are conducted with eleven state-of-the-art domain adaptation methods. Experimental results demonstrate that our proposed method achieves superior performances on the six open-set cross-domain scenarios in both scene classification and retrieval. Especially, our method improves the overall classification accuracies by 9.72% to 24.06% and improve mean average retrieval precisions by 8.06% to 16.21% on the complex UCMD (source domain) → NWPU (target domain) scenario, compared with the other eleven state-of-the-art methods.

Pseudo-Label Noise Prevention, Suppression and Softening for Unsupervised Person Re-Identification

Unsupervised person re-identification (ReID), including fully unsupervised ReID and unsupervised domain adaptive ReID, remains a challenge for the fields of biometrics and computer vision due to its difficulty in learning with unlabeled target domain data. Existing state-of-the-art methods, most of which generate pseudo-labels via unsupervised clustering for model optimization, are inevitably hampered by the under-explored problem of pseudo-label noise. Motivated by this, we propose a novel joint framework termed pseudo-label Noise Prevention, Suppression, and Softening (NPSS) for unsupervised person re-identification. Instead of refining generated label noise after clustering as many existing methods do, we start solving this issue from the source of pseudo-label noise by proposing a new Dynamic Camera-Adaptive Clustering (DCAC), which dynamically involves camera information to prevent noise caused by cross-camera variance, thus improving their quality during clustering. Moreover, we propose an Online Domain Union (ODU) mechanism for the classification model learning on the target domain via involving source domain data with their ground-truth labels, which effectively suppresses the indelible noisy pseudo-labels. Furthermore, we present the Self-Consistency Constraint (SCC) to soften the label noise in a single model with reduced computation and network parameter cost, which achieves intra-sample knowledge ensembling with our global-local SCC and cross-sample knowledge ensembling with our inter-instance SCC. Experiments demonstrate the effectiveness of our method as it surpasses state-of-the-art methods by a large margin on Market-1501, DukeMTMC-ReID, and MSMT17 benchmarks. The code is available at https://github.com/hjwang-824/NPSS.

DualDA-Net: Dual-Head Rectification for Cross-Domain Object Detection of Remote Sensing

Object detectors of remote sensing (RS) imagery with deep learning have become increasingly popular and rely heavily on extensive labeled data. The source-only detectors, which are trained on massive labeled data in a source domain, in some cases fail to get satisfactory performance on a target domain due to the domain shift. To alleviate the domain shift, popular approaches consider feature distribution alignment, but the target domain with massive unlabeled data is under-utilized. Some methods use the source-only model to generate pseudo labels for target domain data, but the variation of different remote sensing scenarios produces the domain shift, which is injected directly into the pseudo labels. Therefore, We propose the Dual-head rectification Domain Adaptation network (DualDA-Net) to alleviate the domain shift and exploit the potential of unlabeled target domain data. DualDA-Net cooperates the coarse-to-fine consistency alignment (CCA) with dual-head co-training (DHCT) to align the distribution and generate pseudo labels progressively. Specifically, the CCA focuses on source and target domain feature distribution alignment via coarse-to-fine consistency alignment on multi-level features. Moreover, the DHCT with dual detection heads is deployed in the teacher-student framework, where one of the heads complements the other with high-quality predictions to rectify the pseudo labels as supervision and alleviate the biased information. Sufficient experiments have been conducted on several domain adaptation settings. The experimental results demonstrate that our DualDA-Net achieves success in the target domain for cross domain object detection of RS imagery.

CSF: Closed-mask-guided semantic fusion method for semantic perception of unknown scenes

• A mask-guided semantic fusion method for optimizing the results of semantic segmentation. • A closed mask generation (CEG) module for post-processing of the edge detection algorithm. • A semantic confidence fusion (SCF) module to fuse the semantic segmentation results. • This method does not need to obtain the target data in advance. • This method has been verified in two scenarios. Though many high-precision semantic segmentation models have been proposed, how to improve the generalization ability of these models is still an urgent problem. Recently, a great number of unsupervised domain adaptation (UDA) algorithms around domain adaptation problems have been studied in semantic segmentation. These methods require labeled source domain data and unlabeled target domain data. In this paper, we propose a closed-mask-guided semantic fusion method (CSF) to improve the semantic segmentation results of unknown scenes, where the target domain data is not obtained in advance. First, a Closed Mask Generation (CMG) module is designed to convert the edge detection result into a mask that can segment the image into several image blocks. Then, a Semantic Confidence Fusion (SCF) module based on information entropy and voting method is introduced, which can select reliable semantic segmentation results for each image block by comparing the confidence of several semantic segmentation networks . In addition, the experimental results on both KITTI and COCO Stuff datasets validate the effectiveness of this method. The code is publicly available at https://github.com/tryhere/CSF .

Multibranch Adversarial Regression for Domain Adaptative Hand Pose Estimation

Although hand pose estimation has achieved a great success in recent years, there are still challenges with RGB-based estimation tasks, the most significant of which is the absence of labeled training data. At present, the synthetic dataset has plenty of images with accurate annotation, but the difference from real-world datasets affects generalization. Therefore, a transfer learning strategy, which tries to transfer knowledge from a labeled source domain to an unlabeled target domain, is a frequent solution. Existing methods such as mean-teacher, Cyclegan, and MCD will train models with the help of some easily accessible domains such as synthetic data. However, these methods are not guaranteed to operate well in real-world settings due to the domain shift. In this paper, we design a new unsupervised domain adaptation method named Multi-branch Adversarial Regressors (MarsDA) in hand pose estimation, where it could be better for feature migration. Specifically, we first generate pseudo-labels for unlabeled target domain data. Then, the new adversarial training loss between multiple regression branches we designed for hand pose estimation is introduced to narrow the domain gap. In this way, our model can reduce the noise of pseudo labels caused by the domain gap and improve the accuracy of pseudo labels. We evaluate our method on two publicly available real-world datasets, H3D and STB. Experimental results show that our method outperforms existing methods by a large margin.

Edge-preserving Image Synthesis for Unsupervised Domain Adaptation in Medical Image Segmentation.

Domain Adaptation is a technique to address the lack of massive amounts of labeled data in different application domains. Unsupervised domain adaptation is the process of adapting a model to an unseen target dataset using solely labeled source data and unlabeled target domain data. Though many image-spaces domain adaptation methods have been proposed to capture pixel-level domain-shift, such techniques may fail to maintain high-level semantic information for the segmentation task. For the case of biomedical images, fine details such as blood vessels can be lost during the image transformation operations between domains. In this work, we propose a model that adapts between domains using cycle-consistent loss while maintaining edge details of the original images by enforcing an edge-based loss during the adaptation process. We demonstrate the effectiveness of our algorithm by comparing it to other approaches on two eye fundus vessels segmentation datasets. We achieve 3.1 % increment in Dice score compared to the SOTA and ∼ 7.02% increment compared to a vanilla CycleGAN implementation. Clinical relevance- The proposed adaptation scheme can provide better performance on unseen data for semantic segmentation, which is widely applied in computer-aided diagnosis. Such robust performance can reduce the reliance of a large amount of labeled data, which is a common problem in the medical domain.

Application of domain-adaptive convolutional variational autoencoder for stress-state prediction

Applying data-driven methods such as deep learning in material mechanics is challenging because producing a sufficiently large, labeled dataset is costly resource-wise. This paper outlines a new approach to overcoming this difficulty by transferring knowledge from a source domain of finite-element-analysis data to a target domain of real-world test-specimen images so that a model capable of accurate and robust predictions in both domains may be constructed. To achieve this transfer of knowledge, discrepancy-based unsupervised domain adaptation is adopted into a convolutional variational autoencoder structure. To evaluate the proposed approach, a four-point bending experiment was conducted on 6061 aluminum alloy and 316 stainless steel to produce 550 unlabeled target-domain data images. The same bending situation was analyzed using the finite-element method implemented in the commercial software package ABAQUS to produce 6000 labeled, source-domain data images. The proposed domain-adaptive convolutional variational autoencoder was trained using the maximum mean discrepancy method on the target- and the source-domain data. The predictions using the domain-adapted convolutional variational autoencoder were relatively more accurate than those using the model trained only on the source domain. It is expected that the proposed approach can address the scarcity of labeled data in various applications of material mechanics and provide a base technology for the development of various data-driven approaches.

Contrastive transformer based domain adaptation for multi-source cross-domain sentiment classification

Cross-domain sentiment classification aims to predict the sentiment tendency in unlabeled target domain data using labeled source-domain data. The wide range of data sources has motivated research into multi-source cross-domain sentiment classification tasks. Conventional domain adaptation methods focus on reducing the domain difference between the source and target domains to realize sentiment migration, which ignores the selection of effective sources and fails to deal with negative transfer, leading to limited performance. To address these problems, we propose a contrastive transformer-based domain adaptation (CTDA) method, which not only develops a multi-source domain selection strategy, but also improves the problem of negative transfer from the perspective of data quality. Specifically, the proposed CTDA includes four stages: (1) designing a mixed selector to weight all related sources or pick out the Top-K sources according to the spatial similarity between both domains, (2) building an adaptor to extract domain-invariant information of features by minimizing the Wasserstein distance between both domains, (3) constructing a discriminator to capture the domain-private information of features by contrastive learning, and (4) performing a weighted classifier to predict the sentiment tendency of the target domain according to multiple trained source classifiers. Extensive experiments were performed on two public benchmarks, and the results demonstrated that our CTDA model significantly outperforms state-of-the-art approaches.

A Transfer Learning Framework with a One-Dimensional Deep Subdomain Adaptation Network for Bearing Fault Diagnosis under Different Working Conditions.

Accurate and fast rolling bearing fault diagnosis is required for the normal operation of rotating machinery and equipment. Although deep learning methods have achieved excellent results for rolling bearing fault diagnosis, the performance of most methods declines sharply when the working conditions change. To address this issue, we propose a one-dimensional lightweight deep subdomain adaptation network (1D-LDSAN) for faster and more accurate rolling bearing fault diagnosis. The framework uses a one-dimensional lightweight convolutional neural network backbone for the rapid extraction of advanced features from raw vibration signals. The local maximum mean discrepancy (LMMD) is employed to match the probability distribution between the source domain and the target domain data, and a fully connected neural network is used to identify the fault classes. Bearing data from the Case Western Reserve University (CWRU) datasets were used to validate the performance of the proposed framework under different working conditions. The experimental results show that the classification accuracy for 12 tasks was higher for the 1D-LDSAN than for mainstream transfer learning methods. Moreover, the proposed framework provides satisfactory results when a small proportion of the unlabeled target domain data is used for training.

An Efficient Deep Unsupervised Domain Adaptation for Unknown Malware Detection

As an innovative way of communicating information, the Internet has become an indispensable part of our lives. However, it also facilitates a more widespread attack of malware. With the assistance of modern cryptanalysis, emerging malware having symmetric properties, such as encryption and decryption, pack and unpack, presents new challenges to effective malware detection. Currently, numerous malware detection approaches are based on supervised learning. The biggest challenge is that the existing systems rely on a large amount of labeled data, which is usually difficult to gain. Moreover, since the newly emerging malware has a different data distribution from the original training samples, the detection performance of these systems will degrade along with the emergence of new malware. To solve these problems, we propose an Unsupervised Domain Adaptation (UDA)-based malware detection method by jointly aligning the distribution of known and unknown malware. Firstly, the distribution divergence between the source and target domain is minimized with the help of symmetric adversarial learning to learn shared feature representations. Secondly, to further obtain semantic information of unlabeled target domain data, this paper reduces the class-level distribution divergence by aligning the class center of labeled source and pseudo-labeled target domain data. Finally, we mainly use a residual network with a self-attention mechanism to extract more accurate feature information. A series of experiments are performed on two public datasets. Experimental results illustrate that the proposed approach outperforms the existing detection methods with an accuracy of 95.63% and 95.04% in detecting unknown malware on two datasets, respectively.

Federated Transfer Learning for Bearing Fault Diagnosis With Discrepancy-Based Weighted Federated Averaging

Generally, high performance of deep learning (DL)-based machinery fault diagnosis methods relies on abundant labeled fault samples under various working conditions, while they are usually stored by different users and not communicated with each other due to data privacy protection. Federated learning (FL) is a possible solution, but the traditional federated averaging (FedAvg) algorithm in FL ignores the potential domain shift of different FL participants, which limits its further application. Therefore, a federated transfer learning framework with discrepancy-based weighted federated averaging (D-WFA) is proposed to train the good global diagnosis model collaboratively as well as protect data privacy. First, local labeled source domain data and unlabeled target domain data are utilized to update multiple local models with generalization ability. Then, a maximum mean discrepancy (MMD)-based dynamic weighted averaging algorithm is designed to aggregate the updated local models with automatically learned weight. The proposed D-WFA overcomes the disadvantage of the traditional FedAvg algorithm which assumes all clients have the same contribution in constructing the global model during FL training. Experiment results on a bearing dataset show that the proposed D-WFA outperforms the traditional FedAvg and relative federated transfer learning method, which offers a promising solution in privacy-preserving machine learning for fault diagnosis.

Adaptive Boosting for Domain Adaptation: Toward Robust Predictions in Scene Segmentation.

Domain adaptation is to transfer the shared knowledge learned from the source domain to a new environment, i.e., target domain. One common practice is to train the model on both labeled source-domain data and unlabeled target-domain data. Yet the learned models are usually biased due to the strong supervision of the source domain. Most researchers adopt the early-stopping strategy to prevent over-fitting, but when to stop training remains a challenging problem since the lack of the target-domain validation set. In this paper, we propose one efficient bootstrapping method, called Adaboost Student, explicitly learning complementary models during training and liberating users from empirical early stopping. Adaboost Student combines deep model learning with the conventional training strategy, i.e., adaptive boosting, and enables interactions between learned models and the data sampler. We adopt one adaptive data sampler to progressively facilitate learning on hard samples and aggregate "weak" models to prevent over-fitting. Extensive experiments show that (1) Without the need to worry about the stopping time, AdaBoost Student provides one robust solution by efficient complementary model learning during training. (2) AdaBoost Student is orthogonal to most domain adaptation methods, which can be combined with existing approaches to further improve the state-of-the-art performance. We have achieved competitive results on three widely-used scene segmentation domain adaptation benchmarks.

Semisupervised Heterogeneous Domain Adaptation via Dynamic Joint Correlation Alignment Network for Ship Classification in SAR Imagery

Improving ship classification performance in synthetic aperture radar (SAR) imagery by transferring knowledge from the related domain is a newly emerging research topic. Existing methods follow supervised or unsupervised homogeneous transfer learning techniques with certain restrictions on the use of features (homogeneous rather than heterogeneous) and data (ignoring excavate the potential of unlabeled target domain data), which may hinder further performance improvements. To address these problems, this letter proposes a dynamic joint correlation alignment (DJ-CORAL) network to conduct semi-supervised heterogeneous domain adaptation (HDA). Specifically, DJ-CORAL firstly transforms the heterogeneous features from the source and target domains into a common subspace to eliminate the heterogeneity, then simultaneously performs classifier adaptation and joint marginal and conditional distribution alignment to facilitate the domain shift minimization. Comprehensive experiments validate the superiority of the proposed DJ-CORAL network against state-of-the-art HDA methods. The codes are available at https://github.com/BUCT-RS-ML/DJ-CORAL.

Source data-free domain adaptation for a faster R-CNN

The existing domain adaptive object detection methods often need to carry a large number of source domain samples for domain adaptation, which is not realistic due to GPU limitations, privacy and physical memory in practical applications. To solve this problem, we propose a source data-free domain adaptive object detection method. Only unlabeled target domain data is used to optimize the source domain model so that it can work better in the target domain. Our method takes Faster R-CNN as baseline. Specifically, we first construct global class prototypes which will be updated in batch iteratively. Then based on the global class prototypes, more accurate pseudo-labels are generated for training the target model. In this way, the source and target domains are also implicitly aligned. Our contributions are 1) a prototype guided domain adaptation method which uses prototypes to mine the semantic category information without accessing the source dataset; 2) a scheme of iteratively updating global class prototype which can handle the class and sample imbalances in the training procedure and 3) a more accurate pseudo-label generation method combining semantic information and image information. On multiple public domain adaptive scenarios, our method achieves the state-of-the-art results in terms of accuracy compared with the Faster R-CNN model and some domain adaptive methods with source datasets.

CADA: Multi-scale Collaborative Adversarial Domain Adaptation for unsupervised optic disc and cup segmentation

Recently, deep neural networks have demonstrated comparable and even better performance than board-certified ophthalmologists in well-annotated datasets. However, the diversity of retinal imaging devices poses a significant challenge: domain shift, which leads to performance degradation when applying the deep learning models trained on one domain to new testing domains. In this paper, we propose a domain adaptation framework comprising of multi-scale inputs along with multiple domain adaptors applied hierarchically in both feature and output spaces. The proposed training strategy and novel unsupervised domain adaptation framework, called Collaborative Adversarial Domain Adaptation (CADA), can effectively overcome the domain shift challenge. Multi-scale inputs can reduce the information loss due to the pooling layers used in the network for feature extraction, while our proposed CADA is an interactive paradigm that presents an exquisite collaborative adaptation through both adversarial learning and ensembling weights at different network layers. In particular, to produce a better prediction for the unlabeled target domain data, we simultaneously achieve domain invariance and model generalizability via adversarial learning at multi-scale outputs from different levels of network layers and maintaining an exponential moving average (EMA) of the historical weights during training. Without annotating any sample from the target domain, multiple adversarial losses in encoder and decoder layers guide the extraction of domain-invariant features to confuse the domain classifier. Meanwhile, the ensembling of weights via EMA reduces the uncertainty of adapting multiple discriminator learning. Comprehensive experimental results demonstrate that our CADA model incorporating multi-scale input training can overcome performance degradation and outperform state-of-the-art domain adaptation methods in segmenting retinal optic disc and cup from fundus images stemming from the REFUGE, Drishti-GS, and Rim-One-r3 datasets. The code is available at https://github.com/cswin/CADA.

Spacecraft Intelligent Fault Diagnosis under Variable Working Conditions via Wasserstein Distance-Based Deep Adversarial Transfer Learning

In recent years, deep learning methods which promote the accuracy and efficiency of fault diagnosis task without any extra requirement of artificial feature extraction have elicited the attention of researchers in the field of manufacturing industry as well as aerospace. However, the problems that data in source and target domains usually have different probability distributions because of different working conditions and there are insufficient labeled or even unlabeled data in target domain significantly deteriorate the performance and generalization of deep fault diagnosis models. To address these problems, we propose a novel Wasserstein Generative Adversarial Network with Gradient Penalty- (WGAN-GP-) based deep adversarial transfer learning (WDATL) model in this study, which exploits a domain critic to learn domain invariant feature representations by minimizing the Wasserstein distance between the source and target feature distributions through adversarial training. Moreover, an improved one-dimensional convolutional neural network- (CNN-) based feature extractor which utilizes exponential linear units (ELU) as activation functions and wide kernels is designed to automatically extract the latent features of raw time-series input data. Then, the fault model classifier trained in one working condition (source domain) with sufficient labeled samples could be generalized to diagnose data in other working conditions (target domain) with insufficient labeled samples. Experiments on two open datasets demonstrate that our proposed WDATL model outperforms most of the state-of-the-art approaches on transfer diagnosis tasks under diverse working circumstances.

A novel multi-adversarial cross-domain neural network for bearing fault diagnosis

Recently, deep neural networks have achieved great success in bearing fault diagnosis. Most existing methods are developed under the assumption that the bearing vibration signals are collected under the same machine operating conditions. However, bearing fault diagnosis under cross-domain conditions will suffer from domain shift problems if the neural network is only trained with the source domain data. Moreover, acquiring enough labeled data from the target domain will be expensive and time-consuming. To address the above problems, this paper proposes an end-to-end multi-adversarial cross-domain neural network for bearing fault diagnosis, which takes labeled source domain data and unlabeled target domain data to achieve the cross-domain bearing fault diagnosis under cross-load conditions and cross-machine conditions. The proposed method employs multi-adversarial training to automatically extract the domain-invariant features from source and target domains instead of manually designing features, which combines domain-adversarial learning and mini-max entropy adversarial learning to adversarially reduce the domain discrepancy between the source and target domains and alleviate the class misalignment problem. The results of the cross-load and the cross-machine experiments prove the effectiveness of the proposed method, and the proposed method provides a promising tool for cross-domain bearing fault diagnosis.