Metal artefact reduction (MAR) techniques aim at removing metal-induced noise from clinical images. In Computed Tomography (CT), supervised deep learning approaches have been shown effective but limited in generalisability, as they mostly rely on synthetic data. In Magnetic Resonance Imaging (MRI) instead, no method has yet been introduced to correct the susceptibility artefact, still present even in MAR-specific acquisitions. In this work, we hypothesise that a multimodal approach to MAR would improve both CT and MRI. Given their different artefact appearance, their complementary information can compensate for the corrupted signal in either modality. We thus propose an unsupervised deep learning method for multimodal MAR. We introduce the use of Locally Normalised Cross Correlation as a loss term to encourage the fusion of multimodal information. Experiments show that our approach favours a smoother correction in the CT, while promoting signal recovery in the MRI.
Multiple sclerosis lesion activity segmentation is the task of detecting new and enlarging lesions that appeared between a baseline and a follow-up brain MRI scan. While deep learning methods for single-scan lesion segmentation are common, deep learning approaches for lesion activity have only been proposed recently. Here, a two-path architecture processes two 3D MRI volumes from two time points. In this work, we investigate whether extending this problem to full 4D deep learning using a history of MRI volumes and thus an extended baseline can improve performance. For this purpose, we design a recurrent multi-encoder-decoder architecture for processing 4D data. We find that adding more temporal information is beneficial and our proposed architecture outperforms previous approaches with a lesion-wise true positive rate of 0.84 at a lesion-wise false positive rate of 0.19.
Hyperspectral Image (HSI) classification based on deep learning has been an attractive area in recent years. However, as a kind of data-driven algorithm, deep learning method usually requires numerous computational resources and high-quality labelled dataset, while the cost of high-performance computing and data annotation is expensive. In this paper, to reduce dependence on massive calculation and labelled samples, we propose a lightweight network architecture (LiteDenseNet) based on DenseNet for Hyperspectral Image Classification. Inspired by GoogLeNet and PeleeNet, we design a 3D two-way dense layer to capture the local and global features of the input. As convolution is a computationally intensive operation, we introduce group convolution to decrease calculation cost and parameter size further. Thus, the number of parameters and the consumptions of calculation are observably less than contrapositive deep learning methods, which means LiteDenseNet owns simpler architecture and higher efficiency. A series of quantitative experiences on 6 widely used hyperspectral datasets show that the proposed LiteDenseNet obtains the state-of-the-art performance, even though when the absence of labelled samples is severe.
When a user connects to the Internet to fulfill his needs, he often encounters a huge amount of related information. Recommender systems are the techniques for massively filtering information and offering the items that users find them satisfying and interesting. The advances in machine learning methods, especially deep learning, have led to great achievements in recommender systems, although these systems still suffer from challenges such as cold-start and sparsity problems. To solve these problems, context information such as user communication network is usually used. In this paper, we have proposed a novel recommendation method based on Matrix Factorization and graph analysis methods. In addition, we leverage deep Autoencoders to initialize users and items latent factors, and deep embedding method gathers users' latent factors from the user trust graph. The proposed method is implemented on two standard datasets. The experimental results and comparisons demonstrate that the proposed approach is superior to the existing state-of-the-art recommendation methods. Our approach outperforms other comparative methods and achieves great improvements.
Machine learning and in particular deep learning algorithms are the emerging approaches to data analysis. These techniques have transformed traditional data mining-based analysis radically into a learning-based model in which existing data sets along with their cluster labels (i.e., train set) are learned to build a supervised learning model and predict the cluster labels of unseen data (i.e., test set). In particular, deep learning techniques are capable of capturing and learning hidden features in a given data sets and thus building a more accurate prediction model for clustering and labeling problem. However, the major problem is that time series data are often unlabeled and thus supervised learning-based deep learning algorithms cannot be directly adapted to solve the clustering problems for these special and complex types of data sets. To address this problem, this paper introduces a two-stage method for clustering time series data. First, a novel technique is introduced to utilize the characteristics (e.g., volatility) of given time series data in order to create labels and thus be able to transform the problem from unsupervised learning into supervised learning. Second, an autoencoder-based deep learning model is built to learn and model both known and hidden features of time series data along with their created labels to predict the labels of unseen time series data. The paper reports a case study in which financial and stock time series data of selected 70 stock indices are clustered into distinct groups using the introduced two-stage procedure. The results show that the proposed procedure is capable of achieving 87.5\% accuracy in clustering and predicting the labels for unseen time series data.
We introduce a novel approach to automatic unstructured mesh generation using machine learning to predict an optimal finite element mesh for a previously unseen problem. The framework that we have developed is based around training an artificial neural network (ANN) to guide standard mesh generation software, based upon a prediction of the required local mesh density throughout the domain. We describe the training regime that is proposed, based upon the use of \emph{a posteriori} error estimation, and discuss the topologies of the ANNs that we have considered. We then illustrate performance using two standard test problems, a single elliptic partial differential equation (PDE) and a system of PDEs associated with linear elasticity. We demonstrate the effective generation of high quality meshes for arbitrary polygonal geometries and a range of material parameters, using a variety of user-selected error norms.
A bottleneck in any evolutionary art system is aesthetic evaluation. Many different methods have been proposed to automate the evaluation of aesthetics, including measures of symmetry, coherence, complexity, contrast and grouping. The interactive genetic algorithm (IGA) relies on human-in-the-loop, subjective evaluation of aesthetics, but limits possibilities for large search due to user fatigue and small population sizes. In this paper we look at how recent advances in deep learning can assist in automating personal aesthetic judgement. Using a leading artist's computer art dataset, we use dimensionality reduction methods to visualise both genotype and phenotype space in order to support the exploration of new territory in any generative system. Convolutional Neural Networks trained on the user's prior aesthetic evaluations are used to suggest new possibilities similar or between known high quality genotype-phenotype mappings.
Probabilistic password strength meters have been proved to be the most accurate tools to measure password strength. Unfortunately, by construction, they are limited to solely produce an opaque security estimation that fails to fully support the user during the password composition. In the present work, we move the first steps towards cracking the intelligibility barrier of this compelling class of meters. We show that probabilistic password meters inherently own the capability of describing the latent relation occurring between password strength and password structure. In our approach, the security contribution of each character composing a password is disentangled and used to provide explicit fine-grained feedback for the user. Furthermore, unlike existing heuristic constructions, our method is free from any human bias, and, more importantly, its feedback has a probabilistic interpretation. In our contribution: (1) we formulate interpretable probabilistic password strength meters; (2) we describe how they can be implemented via an efficient and lightweight deep learning framework suitable for client-side operability.
Analyzing the ever-increasing volume of posts on social media sites such as Facebook and Twitter requires improved information processing methods for profiling authorship. Document classification is central to this task, but the performance of traditional supervised classifiers has degraded as the volume of social media has increased. This paper addresses this problem in the context of gender detection through ensemble classification that employs multi-model deep learning architectures to generate specialized understanding from different feature spaces.
Hate speech detection is a challenging problem with most of the datasets available in only one language: English. In this paper, we conduct a large scale analysis of multilingual hate speech in 9 languages from 16 different sources. We observe that in low resource setting, simple models such as LASER embedding with logistic regression performs the best, while in high resource setting BERT based models perform better. In case of zero-shot classification, languages such as Italian and Portuguese achieve good results. Our proposed framework could be used as an efficient solution for low-resource languages. These models could also act as good baselines for future multilingual hate speech detection tasks. We have made our code and experimental settings public for other researchers at https://github.com/punyajoy/DE-LIMIT.
This paper aims to handle the joint transmitter and noncoherent receiver design for multiuser multiple-input multiple-output (MU-MIMO) systems through deep learning. Given the deep neural network (DNN) based noncoherent receiver, the novelty of this work mainly lies in the multiuser waveform design at the transmitter side. According to the signal format, the proposed deep learning solutions can be divided into two groups. One group is called pilot-aided waveform, where the information-bearing symbols are time-multiplexed with the pilot symbols. The other is called learning-based waveform, where the multiuser waveform is partially or even completely designed by deep learning algorithms. Specifically, if the information-bearing symbols are directly embedded in the waveform, it is called systematic waveform. Otherwise, it is called non-systematic waveform, where no artificial design is involved. Simulation results show that the pilot-aided waveform design outperforms the conventional zero forcing receiver with least squares (LS) channel estimation on small-size MU-MIMO systems. By exploiting the time-domain degrees of freedom (DoF), the learning-based waveform design further improves the detection performance by at least 5 dB at high signal-to-noise ratio (SNR) range. Moreover, it is found that the traditional weight initialization method might cause a training imbalance among different users in the learning-based waveform design. To tackle this issue, a novel weight initialization method is proposed which provides a balanced convergence performance with no complexity penalty.
Autonomous vehicles were experiencing rapid development in the past few years. However, achieving full autonomy is not a trivial task, due to the nature of the complex and dynamic driving environment. Therefore, autonomous vehicles are equipped with a suite of different sensors to ensure robust, accurate environmental perception. In particular, the camera-LiDAR fusion is becoming an emerging research theme. However, so far there has been no critical review that focuses on deep-learning-based camera-LiDAR fusion methods. To bridge this gap and motivate future research, this paper devotes to review recent deep-learning-based data fusion approaches that leverage both image and point cloud. This review gives a brief overview of deep learning on image and point cloud data processing. Followed by in-depth reviews of camera-LiDAR fusion methods in depth completion, object detection, semantic segmentation, tracking and online cross-sensor calibration, which are organized based on their respective fusion levels. Furthermore, we compare these methods on publicly available datasets. Finally, we identified gaps and over-looked challenges between current academic researches and real-world applications. Based on these observations, we provide our insights and point out promising research directions.
The ability to predict, anticipate and reason about future outcomes is a key component of intelligent decision-making systems. In light of the success of deep learning in computer vision, deep-learning-based video prediction emerged as a promising research direction. Defined as a self-supervised learning task, video prediction represents a suitable framework for representation learning, as it demonstrated potential capabilities for extracting meaningful representations of the underlying patterns in natural videos. Motivated by the increasing interest in this task, we provide a review on the deep learning methods for prediction in video sequences. We firstly define the video prediction fundamentals, as well as mandatory background concepts and the most used datasets. Next, we carefully analyze existing video prediction models organized according to a proposed taxonomy, highlighting their contributions and their significance in the field. The summary of the datasets and methods is accompanied with experimental results that facilitate the assessment of the state of the art on a quantitative basis. The paper is summarized by drawing some general conclusions, identifying open research challenges and by pointing out future research directions.
In-database machine learning has been very popular, almost being a cliche. However, can we do it the other way around? In this work, we say "yes" by applying plain old SQL to deep learning, in a sense implementing deep learning algorithms with SQL. Most deep learning frameworks, as well as generic machine learning ones, share a de facto standard of multidimensional array operations, underneath fancier infrastructure such as automatic differentiation. As SQL tables can be regarded as generalisations of (multi-dimensional) arrays, we have found a way to express common deep learning operations in SQL, encouraging a different way of thinking and thus potentially novel models. In particular, one of the latest trend in deep learning was the introduction of sparsity in the name of graph convolutional networks, whereas we take sparsity almost for granted in the database world. As both databases and machine learning involve transformation of datasets, we hope this work can inspire further works utilizing the large body of existing wisdom, algorithms and technologies in the database field to advance the state of the art in machine learning, rather than merely integerating machine learning into databases.
Analyzing a huge amount of malware is a major burden for security analysts. Since emerging malware is often a variant of existing malware, automatically classifying malware into known families greatly reduces a part of their burden. Image-based malware classification with deep learning is an attractive approach for its simplicity, versatility, and affinity with the latest technologies. However, the impact of differences in deep learning models and the degree of transfer learning on the classification accuracy of malware variants has not been fully studied. In this paper, we conducted an exhaustive survey of deep learning models using 24 ImageNet pre-trained models and five fine-tuning parameters, totaling 120 combinations, on two platforms. As a result, we found that the highest classification accuracy was obtained by fine-tuning one of the latest deep learning models with a relatively low degree of transfer learning, and we achieved the highest classification accuracy ever in cross-validation on the Malimg and Drebin datasets. We also confirmed that this trend holds true for the recent malware variants using the VirusTotal 2020 Windows and Android datasets. The experimental results suggest that it is effective to periodically explore optimal deep learning models with the latest models and malware datasets by gradually reducing the degree of transfer learning from half.
Deep learning is a state of the art method for a lot of applications. The main issue is that most of the real-time data is highly imbalanced in nature. In order to avoid bias in training, cost-sensitive approach can be used. In this paper, we propose cost-sensitive deep learning based frameworks and the performance of the frameworks is evaluated on three different Cyber Security use cases which are Domain Generation Algorithm (DGA), Electronic mail (Email), and Uniform Resource Locator (URL). Various experiments were performed using cost-insensitive as well as cost-sensitive methods and parameters for both of these methods are set based on hyperparameter tuning. In all experiments, the cost-sensitive deep learning methods performed better than the cost-insensitive approaches. This is mainly due to the reason that cost-sensitive approach gives importance to the classes which have a very less number of samples during training and this helps to learn all the classes in a more efficient manner.
As the Internet is growing rapidly these years, the variant of malicious software, which often referred to as malware, has become one of the major and serious threats to Internet users. The dramatic increase of malware has led to a research area of not only using cutting edge machine learning techniques classify malware into their known families, moreover, recognize the unknown ones, which can be related to Open Set Recognition (OSR) problem in machine learning. Recent machine learning works have shed light on Open Set Recognition (OSR) from different scenarios. Under the situation of missing unknown training samples, the OSR system should not only correctly classify the known classes, but also recognize the unknown class. This survey provides an overview of different deep learning techniques, a discussion of OSR and graph representation solutions and an introduction of malware classification systems.
This paper presents and explores a robust deep learning framework for auscultation analysis. This aims to classify anomalies in respiratory cycles and detect disease, from respiratory sound recordings. The framework begins with front-end feature extraction that transforms input sound into a spectrogram representation. Then, a back-end deep learning network is used to classify the spectrogram features into categories of respiratory anomaly cycles or diseases. Experiments, conducted over the ICBHI benchmark dataset of respiratory sounds, confirm three main contributions towards respiratory-sound analysis. Firstly, we carry out an extensive exploration of the effect of spectrogram type, spectral-time resolution, overlapped/non-overlapped windows, and data augmentation on final prediction accuracy. This leads us to propose a novel deep learning system, built on the proposed framework, which outperforms current state-of-the-art methods. Finally, we apply a Teacher-Student scheme to achieve a trade-off between model performance and model complexity which additionally helps to increase the potential of the proposed framework for building real-time applications.
Deep learning based models have surpassed classical machine learning based approaches in various text classification tasks, including sentiment analysis, news categorization, question answering, and natural language inference. In this paper, we provide a comprehensive review of more than 150 deep learning based models for text classification developed in recent years, and discuss their technical contributions, similarities, and strengths. We also provide a summary of more than 40 popular datasets widely used for text classification. Finally, we provide a quantitative analysis of the performance of different deep learning models on popular benchmarks, and discuss future research directions.
The current wave of deep learning (the hyper-vitamined return of artificial neural networks) applies not only to traditional statistical machine learning tasks: prediction and classification (e.g., for weather prediction and pattern recognition), but has already conquered other areas, such as translation. A growing area of application is the generation of creative content, notably the case of music, the topic of this paper. The motivation is in using the capacity of modern deep learning techniques to automatically learn musical styles from arbitrary musical corpora and then to generate musical samples from the estimated distribution, with some degree of control over the generation. This paper provides a tutorial on music generation based on deep learning techniques. After a short introduction to the topic illustrated by a recent exemple, the paper analyzes some early works from the late 1980s using artificial neural networks for music generation and how their pioneering contributions have prefigured current techniques. Then, we introduce some conceptual framework to analyze the various concepts and dimensions involved. Various examples of recent systems are introduced and analyzed to illustrate the variety of concerns and of techniques.