COVID-19 classification using chest Computed Tomography (CT) has been found pragmatically useful by several studies. Due to the lack of annotated samples, these studies recommend transfer learning and explore the choices of pre-trained models and data augmentation. However, it is still unknown if there are better strategies than vanilla transfer learning for more accurate COVID-19 classification with limited CT data. This paper provides an affirmative answer, devising a novel `model' augmentation technique that allows a considerable performance boost to transfer learning for the task. Our method systematically reduces the distributional shift between the source and target domains and considers augmenting deep learning with complementary representation learning techniques. We establish the efficacy of our method with publicly available datasets and models, along with identifying contrasting observations in the previous studies.

Micro-expression can reflect people's real emotions. Recognizing micro-expressions is difficult because they are small motions and have a short duration. As the research is deepening into micro-expression recognition, many effective features and methods have been proposed. To determine which direction of movement feature is easier for distinguishing micro-expressions, this paper selects 18 directions (including three types of horizontal, vertical and oblique movements) and proposes a new low-dimensional feature called the Histogram of Single Direction Gradient (HSDG) to study this topic. In this paper, HSDG in every direction is concatenated with LBP-TOP to obtain the LBP with Single Direction Gradient (LBP-SDG) and analyze which direction of movement feature is more discriminative for micro-expression recognition. As with some existing work, Euler Video Magnification (EVM) is employed as a preprocessing step. The experiments on the CASME II and SMIC-HS databases summarize the effective and optimal directions and demonstrate that HSDG in an optimal direction is discriminative, and the corresponding LBP-SDG achieves state-of-the-art performance using EVM.

Recent promises of generative deep learning lately brought interest to its potential uses in neural engineering. In this paper we firstly review recently emerging studies on generating artificial electroencephalography (EEG) signals with deep neural networks. Subsequently, we present our feasibility experiments on generating condition-specific multichannel EEG signals using conditional variational autoencoders. By manipulating real resting-state EEG epochs, we present an approach to synthetically generate time-series multichannel signals that show spectro-temporal EEG patterns which are expected to be observed during distinct motor imagery conditions.

Recent advances in Reinforcement Learning (RL) have surpassed human-level performance in many simulated environments. However, existing reinforcement learning techniques are incapable of explicitly incorporating already known domain-specific knowledge into the learning process. Therefore, the agents have to explore and learn the domain knowledge independently through a trial and error approach, which consumes both time and resources to make valid responses. Hence, we adapt the Deep Deterministic Policy Gradient (DDPG) algorithm to incorporate an adviser, which allows integrating domain knowledge in the form of pre-learned policies or pre-defined relationships to enhance the agent's learning process. Our experiments on OpenAi Gym benchmark tasks show that integrating domain knowledge through advisers expedites the learning and improves the policy towards better optima.

In recent years, physiological signal based authentication has shown great promises,for its inherent robustness against forgery. Electrocardiogram (ECG) signal, being the most widely studied biosignal, has also received the highest level of attention in this regard. It has been proven with numerous studies that by analyzing ECG signals from different persons, it is possible to identify them, with acceptable accuracy. In this work, we present, EDITH, a deep learning-based framework for ECG biometrics authentication system. Moreover, we hypothesize and demonstrate that Siamese architectures can be used over typical distance metrics for improved performance. We have evaluated EDITH using 4 commonly used datasets and outperformed the prior works using less number of beats. EDITH performs competitively using just a single heartbeat (96-99.75% accuracy) and can be further enhanced by fusing multiple beats (100% accuracy from 3 to 6 beats). Furthermore, the proposed Siamese architecture manages to reduce the identity verification Equal Error Rate (EER) to 1.29%. A limited case study of EDITH with real-world experimental data also suggests its potential as a practical authentication system.

The success of modern deep learning algorithms for image segmentation heavily depends on the availability of large datasets with clean pixel-level annotations (masks), where the objects of interest are accurately delineated. Lack of time and expertise during data annotation leads to incorrect boundaries and label noise. It is known that deep convolutional neural networks (DCNNs) can memorize even completely random labels, resulting in poor accuracy. We propose a framework to train binary segmentation DCNNs using sets of unreliable pixel-level annotations. Erroneously labeled pixels are identified based on the estimated aleatoric uncertainty of the segmentation and are relabeled to the true value.

We implement stacked denoising autoencoders, a class of neural networks that are capable of learning powerful representations of high dimensional data. We describe stochastic gradient descent for unsupervised training of autoencoders, as well as a novel genetic algorithm based approach that makes use of gradient information. We analyze the performance of both optimization algorithms and also the representation learning ability of the autoencoder when it is trained on standard image classification datasets. The weight matrix of the decoding stage is the transpose of the weight matrix of the encoding stage. Autoencoders are a method for performing representation learning, an unsupervised pretraining process during which a more useful representation of the input data is automatically determined. Representation learning is important in machine learning since "the performance of machine learning methods is heavily dependent on the choice of data representation (or features) in which they are applied" [1]. For many supervised classification tasks, the high dimensionality of the input data means that the classifier requires an enormous number of training examples in order to generalize well and not overfit. Autoencoders are one such representation learning tool. An autoencoder is a neural network with a single hidden layer and where the output layer and the input layer have the same size. Then we have a neural network as shown in Figure 1. The weight matrix of the decoding stage is the transpose of weight matrix of the encoding stage in order to reduce the number of parameters to learn. After an autoencoder is trained, its decoding stage is discarded and the encoding stage is used to transform the training input examples as a preprocessing step.

U-Net based convolutional neural networks with deep feature representation and skip-connections have significantly boosted the performance of medical image segmentation. In this paper, we study the more challenging problem of improving efficiency in modeling global contexts without losing localization ability for low-level details. TransFuse, a novel two-branch architecture is proposed, which combines Transformers and CNNs in a parallel style. With TransFuse, both global dependency and low-level spatial details can be efficiently captured in a much shallower manner. Besides, a novel fusion technique - BiFusion module is proposed to fuse the multi-level features from each branch. TransFuse achieves the newest state-of-the-arts on polyp segmentation task, with 20\% fewer parameters and the fastest inference speed at about 98.7 FPS.

Binary neural networks (BNNs) have attracted broad research interest due to their efficient storage and computational ability. Nevertheless, a significant challenge of BNNs lies in handling discrete constraints while ensuring bit entropy maximization, which typically makes their weight optimization very difficult. Existing methods relax the learning using the sign function, which simply encodes positive weights into +1s, and -1s otherwise. Alternatively, we formulate an angle alignment objective to constrain the weight binarization to {0,+1} to solve the challenge. In this paper, we show that our weight binarization provides an analytical solution by encoding high-magnitude weights into +1s, and 0s otherwise. Therefore, a high-quality discrete solution is established in a computationally efficient manner without the sign function. We prove that the learned weights of binarized networks roughly follow a Laplacian distribution that does not allow entropy maximization, and further demonstrate that it can be effectively solved by simply removing the $\ell_2$ regularization during network training. Our method, dubbed sign-to-magnitude network binarization (SiMaN), is evaluated on CIFAR-10 and ImageNet, demonstrating its superiority over the sign-based state-of-the-arts. Code is at https://github.com/lmbxmu/SiMaN.

In ophthalmology, early fundus screening is an economic and effective way to prevent blindness caused by ophthalmic diseases. Clinically, due to the lack of medical resources, manual diagnosis is time-consuming and may delay the condition. With the development of deep learning, some researches on ophthalmic diseases have achieved good results, however, most of them are just based on one disease. During fundus screening, ophthalmologists usually give diagnoses of multi-disease on binocular fundus image, so we release a dataset with 8 diseases to meet the real medical scene, which contains 10,000 fundus images from both eyes of 5,000 patients. We did some benchmark experiments on it through some state-of-the-art deep neural networks. We found simply increasing the scale of network cannot bring good results for multi-disease classification, and a well-structured feature fusion method combines characteristics of multi-disease is needed. Through this work, we hope to advance the research of related fields.