Domain Disentangled Meta-Learning

Recent neural architectures in named entity recognition (NER) have yielded state-of-the-art performance on single domain data such as newswires. However, they still suffer from (i) requiring massive amounts of training data to avoid overfitting; (ii) huge performance degradation when there is a domain shift in the data distribution between training and testing. In this paper, we investigate the problem of domain adaptation for NER under homogeneous and heterogeneous settings. We propose MetaNER, a novel meta-learning approach for domain adaptation in NER. Specifically, MetaNER incorporates meta-learning and adversarial training strategies to encourage robust, general and transferable representations for sequence labeling. The key advantage of MetaNER is that it is capable of adapting to new unseen domains with a small amount of annotated data from those domains. We extensively evaluate MetaNER on multiple datasets under homogeneous and heterogeneous settings. The experimental results show that MetaNER achieves state-of-the-art performance against eight baselines. Impressively, MetaNER surpasses the in-domain performance using only 16.17% and 34.76% of target domain data on average for homogeneous and heterogeneous settings, respectively.

Generalising deep models to new data from new centres (termed here domains) remains a challenge. This is largely attributed to shifts in data statistics (domain shifts) between source and unseen domains. Recently, gradient-based meta-learning approaches where the training data are split into meta-train and meta-test sets to simulate and handle the domain shifts during training have shown improved generalisation performance. However, the current fully supervised meta-learning approaches are not scalable for medical image segmentation, where large effort is required to create pixel-wise annotations. Meanwhile, in a low data regime, the simulated domain shifts may not approximate the true domain shifts well across source and unseen domains. To address this problem, we propose a novel semi-supervised meta-learning framework with disentanglement. We explicitly model the representations related to domain shifts. Disentangling the representations and combining them to reconstruct the input image allows unlabeled data to be used to better approximate the true domain shifts for meta-learning. Hence, the model can achieve better generalisation performance, especially when there is a limited amount of labeled data. Experiments show that the proposed method is robust on different segmentation tasks and achieves state-of-the-art generalisation performance on two public benchmarks.

Machine learning algorithms with empirical risk minimization are vulnerable under distributional shifts due to the greedy adoption of all the correlations found in training data. Recently, there are robust learning methods aiming at this problem by minimizing the worst-case risk over an uncertainty set. However, they equally treat all covariates to form the decision sets regardless of the stability of their correlations with the target, resulting in the overwhelmingly large set and low confidence of the learner. In this paper, we propose Stable Adversarial Learning (SAL) algorithm that leverages heterogeneous data sources to construct a more practical uncertainty set and conduct differentiated robustness optimization, where covariates are differentiated according to the stability of their correlations with the target. We theoretically show that our method is tractable for stochastic gradient-based optimization and provide the performance guarantees for our method. Empirical studies on both simulation and real datasets validate the effectiveness of our method in terms of uniformly good performance across unknown distributional shifts.

Unsupervised domain adaptation (UDA) techniques, extensively studied in hyperspectral image (HSI) classification, aim to use labeled source domain data and unlabeled target domain data to learn domain invariant features for cross-scene classification. Compared to natural images, numerous spectral bands of HSIs provide abundant semantic information, but they also increase the domain shift significantly. In most existing methods, both explicit alignment and implicit alignment simply align feature distribution, ignoring domain information in the spectrum. We noted that when the spectral channel between source and target domains is distinguished obviously, the transfer performance of these methods tends to deteriorate. Additionally, their performance fluctuates greatly owing to the varying domain shifts across various datasets. To address these problems, a novel shift-sensitive spatial-spectral disentangling learning (S4DL) approach is proposed. In S4DL, gradient-guided spatial-spectral decomposition (GSSD) is designed to separate domain-specific and domain-invariant representations by generating tailored masks under the guidance of the gradient from domain classification. A shift-sensitive adaptive monitor is defined to adjust the intensity of disentangling according to the magnitude of domain shift. Furthermore, a reversible neural network is constructed to retain domain information that lies not only in semantic but also the shallow-level detailed information. Extensive experimental results on several cross-scene HSI datasets consistently verified that S4DL is better than the state-of-the-art UDA methods. Our source code will be available at https://github.com/xdu-jjgs/IEEE_TNNLS_S4DL.

Existing domain generalization (DG) methods typically rely on source domain data for model training and unknown target domain data to evaluate generalization ability. However, collecting target domain data is challenging because it must be out-of-distribution from the source domain data. Thus, developing Target-Free Evaluation (TFEval) methods is essential to assess the generalization performance of deep models without requiring unknown target domain data, relying only on trained data. In this paper, we estimate generalization performance for TFEval by training a regression predictor. The predictor takes model representations as input and predicts generalization performance. Inspired by the observation that Batch Normalization (BN) parameters strongly reflect domain/distribution information, we extract model representations by calculating BN parameter differences between the trained model and its proxy models. The proxy models are trained on a significantly smaller subset of the source domain data. We construct multiple TFEval datasets by collecting numerous deep models, and experimental results on these datasets demonstrate that our evaluation method effectively predicts the generalization performance of deep models.

A unified framework for visual domain adaptation with covariance matching

Abstract-With the fast advance of mobile chips technologies, a node in a smart camera network can afford sophisticated processing via on-board multicore CPUs and GPUs, e.g., face detection. The performance of a general purpose face detector, however, may degrade seriously under specific situations with unexpected challenges such as facial coverage or bad illumination. This degradation is due to the difference of probability distributions between training data and testing data, known as source data domain and target data domain, respectively. To better adapt a smart camera network to a specific situation, some form of target domain adaptation is needed, which usually requires both the source domain data and as much as possible target domain data at each node, which may strain storage capacity and bandwidth. In this paper, we propose an adaptation method which fuses the source specific hypotheses (SSHs) and target specific hypotheses (TSHs) - requiring only a pre-trained face detector and a few target data to be shared by peer-to-peer communications, thus relieving the storage and bandwidth constraints. The method uses the accuracy-regularization objective as the adaptation model, to fuse SSHs and TSHs, and tries to minimize the misclassification error on target data. With an existing frontal face detector, we conduct experiments to verify our algorithm, covering cases of video surveillance, extreme pose challenge, and different illumination spectra. Significant performance gains are observed with only dozens of target data in all the experiments, demonstrating the effectiveness of the proposed adaptation model. Therefore, the proposed adaption can be applied to a smart camera network with peer-to-peer communications to improve the network's overall face detection performance.

Center-aligned domain adaptation network for image classification

While pre-trained models in the general domain have proliferated, existing methods for transferring these models to specific domains often depend on source domain data for distribution alignment and are typically tailored for single tasks. We propose a source data-free approach, Fourier Self-Adaptation (FSA), which effectively adapts general models to a wide range of specific domains. Our method leverages the distinct properties of Fourier phase and amplitude: phase contains high-level structural and positional information, which is less affected by domain shifts, while amplitude contains details and brightness information, which is more affected by domain shifts. FSA adjusts the image distribution by initializing a trainable adaptive image from a normal distribution. It then interpolates the amplitude of the target domain image with that of the adaptive image, where the interpolation ratio is dynamically controlled by learnable weight and bias. During training, the model captures advanced phase information of the target image and refines the data distribution through amplitude interpolation. Additionally, a dual regularization loss constrains the model representation, encouraging it to focus on the intrinsic relationships of the target domain data while discarding irrelevant knowledge. We evaluate FSA using general pre-trained models on 11 unimodal image classification datasets and 6 multimodal visual question answering datasets, covering specific domains such as radiology, pathology, remote sensing, and art. Our method consistently achieves state-of-the-art performance across multiple datasets, with performance improvements ranging from 1% to 8% compared to basic pre-trained models. Source code are available at https://github.com/Alivelei/FSA.

Neural ordinary differential equations (neural ODEs) have emerged as a novel network architecture that bridges dynamical systems and deep learning. However, the gradient obtained with the continuous adjoint method in the vanilla neural ODE is not reverse-accurate. Other approaches suffer either from an excessive memory requirement due to deep computational graphs or from limited choices for the time integration scheme, hampering their application to large-scale complex dynamical systems. To achieve accurate gradients without compromising memory efficiency and flexibility, we present a new neural ODE framework, PNODE, based on high-level discrete adjoint algorithmic differentiation. By leveraging discrete adjoint time integrators and advanced checkpointing strategies tailored for these integrators, PNODE can provide a balance between memory and computational costs, while computing the gradients consistently and accurately. We provide an open-source implementation based on PyTorch and PETSc, one of the most commonly used portable, scalable scientific computing libraries. We demonstrate the performance through extensive numerical experiments on image classification and continuous normalizing flow problems. We show that PNODE achieves the highest memory efficiency when compared with other reverse-accurate methods. On the image classification problems, PNODE is up to two times faster than the vanilla neural ODE and up to 2.3 times faster than the best existing reverse-accurate method. We also show that PNODE enables the use of the implicit time integration methods that are needed for stiff dynamical systems.

The use of Convolutional Neural Networks (CNNs) to solve Domain Adaptation (DA) image classification problems in the context of remote sensing has proven to provide good results but at high computational cost. To avoid this problem, a deep learning network for DA in remote sensing hyperspectral images called TCANet is proposed. As a standard CNN, TCANet consists of several stages built based on convolutional filters that operate on patches of the hyperspectral image. Unlike the former, the coefficients of the filter are obtained through Transfer Component Analysis (TCA). This approach has two advantages: firstly, TCANet does not require training based on backpropagation, since TCA is itself a learning method that obtains the filter coefficients directly from the input data. Second, DA is performed on the fly since TCA, in addition to performing dimensional reduction, obtains components that minimize the difference in distributions of data in the different domains corresponding to the source and target images. To build an operating scheme, TCANet includes an initial stage that exploits the spatial information by providing patches around each sample as input data to the network. An output stage performing feature extraction that introduces sufficient invariance and robustness in the final features is also included. Since TCA is sensitive to normalization, to reduce the difference between source and target domains, a previous unsupervised domain shift minimization algorithm consisting of applying conditional correlation alignment (CCA) is conditionally applied. The results of a classification scheme based on CCA and TCANet show that the DA technique proposed outperforms other more complex DA techniques.

We address an important and practical problem facing deep-learning-based image classification: class incremental learning with a large domain shift. Most previous efforts on class incremental learning focus on one aspect of the problem, i.e., learning to classify additional new classes (with a little shift). However, in the real world, when new classes are added, the domain changes simultaneously (with a large domain shift). To obtain a model that is robust to these situations, we need to consider incrementally learning not only new labels but also domain-shifted labels. We target a continual and simultaneous shift of class and domain distribution and propose a new incremental learning method named Momentum Contrastive learning enhancing Orthogonality of Negative pairs (MoCo-ON). We employ a momentum encoder framework augmented with rehearsal memory to mitigate the risk of forgetting while leveraging contrastive learning to extract versatile features capable of adapting to the progressively shifting domain. Specifically, when training with a knowledge distillation loss, we introduce a novel supervised contrastive loss designed to closely embed positive pairs of the same class, even in the presence of a substantial domain gap. Additionally, we leverage feature embedding from momentum encoders for exemplar selection, aiming to mitigate the risk of forgetting previously acquired information from earlier tasks. We conduct comprehensive experiments involving inter-domain shifted class incremental learning scenarios using widely adopted datasets commonly employed for studying domain generalization in image classification. Our proposed method consistently outperforms other methods by a significant margin.

The article focuses on studying the effectiveness of two different Hybrid Neural Networks (HNNs) architectures for solving real-world image classification problems. The first approach investigated in the research is a hybridization technique that allows creation of HNN based on a classical neural network by replacing a number of hidden layers of the neural network with a variational quantum circuit, which allows to reduce the complexity of the classical part of the neural network and move part of computations to a quantum device. The second approach is a hybridization technique based on utilizing quanvolutional operations for image processing as the first quantum convolutional layer of the hybrid neural network, thus building a Quanvolutional Neural Network (QNN). QNN leverages quantum phenomena to facilitate feature extraction, enabling the model to achieve higher accuracy metrics than its classical counterpart. The effectiveness of both architectures was tested on several image classification problems. The first one is a classical image classification problem of CIFAR10 images classification, widely used as a benchmark for various imagery-related tasks. Another problem used for the effectiveness study is the problem of geospatial data analysis. The second problem represents a real-world use case where quantum computing utilization can be very fruitful in the future. For studying the effectiveness, several models were assembled: HNN with a quantum device that replaces one of the hidden layers of the neural network, QNN based on quanvolutional operation and utilizes VGG-16 architecture as a classical part of the model, and also an unmodified VGG-16 was used as a reference model. Experiments were conducted to measure the models' key efficiency metrics: maximal accuracy, complexity of a quantum part of the model and complexity of a classical part of the model. The results of the research indicated the feasibility of both approaches for solving both proposed image classification problems. Results were analyzed to outline the advantages and disadvantages of every approach in terms of selected key metrics. Experiments showed that QNN architectures proved to be a feasible and effective solution for critical practical tasks requiring higher levels of model prediction accuracy and, simultaneously, can tolerate higher processing time and significantly increased costs due to a high number of quantum operations required. Also, the results of the experiments indicated that HNN architectures proved to be a feasible solution for time-critical practical tasks that require higher processing speed and can tolerate slightly decreased accuracy of model predictions.

BackgroundHand, foot, and mouth disease (HFMD) has been recognized as one of the leading infectious diseases among children in China, which causes hundreds of annual deaths since 2008. In China, the reports of monthly HFMD cases usually have a delay of 1–2 months due to the time needed for collecting and processing clinical information. This time lag is far from optimal for policymakers making decisions. To alleviate this information gap, this study uses a meta learning framework and combines publicly Internet-based information (Baidu search queries) for real-time estimation of HFMD cases.MethodsWe incorporate Baidu index into modeling to nowcast the monthly HFMD incidences in Guangxi, Zhejiang, Henan provinces and the whole China. We develop a meta learning framework to select appropriate predictive model based on the statistical and time series meta features. Our proposed approach is assessed for the HFMD cases within the time period from July 2015 to June 2016 using multiple evaluation metrics including root mean squared error (RMSE) and correlation coefficient (Corr).ResultsFor the four areas: whole China, Guangxi, Zhejiang, and Henan, our approach is superior to the best competing models, reducing the RMSE by 37, 20, 20, and 30% respectively. Compared with all the alternative predictive methods, our estimates show the strongest correlation with the observations.ConclusionsIn this study, the proposed meta learning method significantly improves the HFMD prediction accuracy, demonstrating that: (1) the Internet-based information offers the possibility for effective HFMD nowcasts; (2) the meta learning approach is capable of adapting to a wide variety of data, and enables selecting appropriate method for improving the nowcasting accuracy.

The concept of Industry 4.0 uses cyber-physical systems and the Internet of Things to create "smart factories" that enable automated and connected production. However, the complex and diverse nature of cyber-physical systems makes them vulnerable to physical threats, where any fault in their physical components can lead to serious consequences, from financial losses to human safety. To reduce the risk of such failures, it is essential to have techniques in place to monitor the condition of these physical components closely. In recent years, data-driven predictive maintenance methods have become common due to the growth in data and advancements in machine learning. This thesis focuses on fault diagnosis, a key part of predictive maintenance, which aims to detect and identify faults in physical components. Incorporating advanced techniques, including machine learning algorithms and sensor data analysis, enhanced fault diagnosis systems' accuracy and efficiency. However, several real-world barriers challenge the effectiveness of these techniques. This research is structured around three primary objectives. The first objective aims to establish an efficient fault diagnosis system that is highly sensitive to fault occurrences while being resilient to uncertainties such as variability in fault patterns and data imbalance. To address variability in fault patterns, we developed a multi-step approach integrating clustering techniques and an adaptive neuro-fuzzy inference system, which significantly improved diagnosis robustness even under noisy conditions. For imbalanced datasets, a two-step method was introduced, combining permutation entropy for fault detection and envelope analysis for fault isolation. This approach was enhanced by an automatic frequency band selection method leveraging wavelet transform and statistical criteria, resulting in high fault diagnosis accuracy and reliability, even with limited labeled faulty data. The second objective focuses on developing fault diagnosis systems that can maintain high performance despite domain shifts and limited training data in both source and target domains. To address the challenge of domain shifts, a transfer learning technique was introduced, and a systematic literature review was conducted. This study bridges a technical gap by providing a comprehensive taxonomy, state-of-the-art methods, and insights into the challenges of applying transfer learning to predictive maintenance. Additionally, a conceptual framework was proposed to explore scenarios for transferring diagnostic knowledge from different source domains to target domains with insufficient training datasets. In a subsequent study, a new diagnostic framework was developed using digital twins and partial transfer learning to address the lack of real training datasets in both source and target domains as well as label space inconsistencies between them. The framework includes a general approach for developing digital twins of real machines to generate synthetic diagnostic data as source domain data. Furthermore, by incorporating a deep learning network and a double-layer attention mechanism, a partial transfer learning diagnostic model was developed to transfer diagnostic knowledge from the digital twin to the real target domain despite label space inconsistencies. The proposed framework was validated through an industrial case study involving rolling bearings in rotating equipment. The third objective focuses on developing an efficient fault diagnosis system capable of maintaining reliability under unseen and evolving working conditions. To achieve this, a novel self-adaptive fault diagnosis system was proposed, incorporating the MAPE-K loop, a domain generalization network model, and digital twin technologies. This self-adaptive fault diagnosis system leverages the MAPE-K loop to adapt in real time to evolving working conditions. Digital twins were employed to continuously generate data aligned with the real system’s operational context, enabling accurate fault diagnosis in dynamic environments. The domain generalization network model enhances generalization to unseen target working conditions by addressing overfitting and feature loss through multi-domain data augmentation, adversarial learning, and domain discrepancy modules. Validation using multiple rotating machinery datasets demonstrated that the developed fault diagnosis system outperformed state-of-the-art methods across various scenarios. In summary, this dissertation proposes a comprehensive and practical approach to addressing variability in fault patterns, imbalance datasets, data scarcity, domain shifts, and unseen and evolving working conditions challenges in existing fault diagnosis systems.