Biblio

Sentiment analytics is the process of applying natural language processing and methods for text-based information to define and extract subjective knowledge of the text. Natural language processing and text classifications can deal with limited corpus data and more attention has been gained by semantic texts and word embedding methods. Deep learning is a powerful method that learns different layers of representations or qualities of information and produces state-of-the-art prediction results. In different applications of sentiment analytics, deep learning methods are used at the sentence, document, and aspect levels. This review paper is based on the main difficulties in the sentiment assessment stage that significantly affect sentiment score, pooling, and polarity detection. The most popular deep learning methods are a Convolution Neural Network and Recurrent Neural Network. Finally, a comparative study is made with a vast literature survey using deep learning models.

Various algorithms and models have been proposed to address text classification tasks; however, they rarely consider incorporating the additional knowledge hidden in class labels. We argue that hidden information in class labels leads to better classification accuracy. In this study, instead of encoding the labels into numerical values, we incorporated the knowledge in the labels into the original model without changing the model architecture. We combined the output of an original classification model with the relatedness calculated based on the embeddings of a sequence and a keyword set. A keyword set is a word set to represent knowledge in the labels. Usually, it is generated from the classes while it could also be customized by the users. The experimental results show that our proposed method achieved statistically significant improvements in text classification tasks. The source code and experimental details of this study can be found on Github11https://github.com/HeroadZ/KiL.

Translation studies is one of the fastest growing interdisciplinary research fields in the world today. Business English is an urgent research direction in the field of translation studies. To some extent, the quality of business English translation directly determines the success or failure of international trade and the economic benefits. On the basis of sequence information encoding and decoding model of LSTM, this paper proposes a strategy combining attention mechanism with bidirectional LSTM model to handle the question of feature extraction of text information. The proposed method reduces the semantic complexity and improves the overall correlation accuracy. The experimental results show its advantages.

Hardware Trojans are modifications made by malicious insiders or third party providers during the design or fabrication phase of the IC (Integrated Circuits) design cycle in a covert manner. These cause catastrophic consequences ranging from manipulating the functionality of individual blocks to disabling the entire chip. Thus, a need for detecting trojans becomes necessary. In this work, we propose a deep learning based approach for detecting trojans in IC chips. In particular, we insert trojans at the circuit-level and generate data by measuring power during normal operation and under attack. Further, we develop deep learning models using Neural networks and Auto-encoders to analyze datasets for outlier detection by profiling the normal behavior and leveraging them to detect anomalies in power consumption. Our approach is generic and non-invasive in that it can be applied to any block without any modifications to the design. Evaluation of the proposed approach shows an accuracy ranging from 92.23% to 99.33% in detecting trojans.

In recent years, the development of deep learning has brought new ideas to internal threat detection. In this paper, three common deep learning algorithms for threat detection are optimized and innovated, and feature embedding, drift detection and sample weighting are introduced into FCNN. Adaptive multi-iteration method is introduced into Support Vector Data Description (SVDD). A dynamic threshold adjustment mechanism is introduced in VAE. In threat detection, three methods are used to detect the abnormal behavior of users, and the intersection of output results is taken as the final threat judgment basis. Experiments on cert r6.2 data set show that this method can significantly reduce the false positive rate.

Cyber security is an important concern in critical infrastructures such as banking and financial organizations, where a number of malicious insiders are involved. These insiders may be existing employees / users present within the organization and causing harm by performing any malicious activity and are commonly known as insider threats. Existing insider threat detection (ITD) methods are based on statistical analysis, machine and deep learning approaches. They monitor and detect malicious user activity based on pre-built rules which fails to detect unforeseen threats. Also, some of these methods require explicit feature engineering which results in high false positives. Apart from this, some methods choose relatively insufficient features and are computationally expensive which affects the classifier's accuracy. Hence, in this paper, a user behaviour based ITD method is presented to overcome the above limitations. It is a conceptually simple and flexible approach based on augmented decision making and anomaly detection. It consists of bi-directional long short term memory (bi-LSTM) for efficient feature extraction. For the purpose of classifying users as "normal" or "malicious", a binary class support vector machine (SVM) is used. CMU-CERT v4.2 dataset is used for testing the proposed method. The performance is evaluated using the following parameters: Accuracy, Precision, Recall, F- Score and AUC-ROC. Test results show that the proposed method outperforms the existing methods.

Computer network and virtual machine security is very essential in today's era. Various architectures have been proposed for network security or prevent malicious access of internal or external users. Various existing systems have already developed to detect malicious activity on victim machines; sometimes any external user creates some malicious behavior and gets unauthorized access of victim machines to such a behavior system considered as malicious activities or Intruder. Numerous machine learning and soft computing techniques design to detect the activities in real-time network log audit data. KKDDCUP99 and NLSKDD most utilized data set to detect the Intruder on benchmark data set. In this paper, we proposed the identification of intruders using machine learning algorithms. Two different techniques have been proposed like a signature with detection and anomaly-based detection. In the experimental analysis, demonstrates SVM, Naïve Bayes and ANN algorithm with various data sets and demonstrate system performance on the real-time network environment.

in this paper, we describe a method for detecting abnormal pedestrians or cars by expressing the behavioral characteristics of pedestrians on a global surveillance map in a video security system using CCTV and patrol robots. This method converts a large amount of video surveillance data into a compressed map shape format to efficiently transmit and process data. By using deep learning auto-encoder and CNN algorithm, pedestrians belonging to the abnormal category can be detected in two steps. In the case of the first-stage abnormal candidate extraction, the normal detection rate was 87.7%, the abnormal detection rate was 88.3%, and in the second stage abnormal candidate filtering, the normal detection rate was 99.8% and the abnormal detection rate was 96.5%.

At present, most smart contract vulnerability detection use manually-defined patterns, which is time-consuming and far from satisfactory. To address this issue, researchers attempt to deploy deep learning techniques for automatic vulnerability detection in smart contracts. Nevertheless, current work mostly relies on a single code representation such as AST (Abstract Syntax Tree) or code tokens to learn vulnerability characteristics, which might lead to incompleteness of learned semantics information. In addition, the number of available vulnerability datasets is also insufficient. To address these limitations, first, we construct a dataset covering most typical types of smart contract vulnerabilities, which can accurately indicate the specific row number where a vulnerability may exist. Second, for each single code representation, we propose a novel way called AFS (AST Fuse program Slicing) to fuse code characteristic information. AFS can fuse the structured information of AST with program slicing information and detect vulnerabilities by learning new vulnerability characteristic information.

In recent years, the complexity of software vulnera-bilities has continued to increase. Manual vulnerability detection methods alone no longer meet the demand. With the rapid development of the deep learning, many neural network models have been widely applied to source code vulnerability detection. The variant of recurrent neural network (RNN), bidirectional Long Short-Term Memory (BiLSTM), has been a popular choice in vulnerability detection. However, is BiLSTM the most suitable choice? To answer this question, we conducted a series of experiments to investigate the effectiveness of different neural network models for source code vulnerability detection. The results shows that the variants of RNN, gated recurrent unit (GRU) and bidirectional GRU, are more capable of detecting source code fragments with mixed vulnerability types. And the concatenated convolutional neural network is more capable of detecting source code fragments of single vulnerability types.

Vulnerabilities in the source code of software are critical issues in the realm of software engineering. Coping with vulnerabilities in software source code is becoming more challenging due to several aspects of complexity and volume. Deep learning has gained popularity throughout the years as a means of addressing such issues. In this paper, we propose an evaluation of vulnerability detection performance on source code representations and evaluate how Machine Learning (ML) strategies can improve them. The structure of our experiment consists of 3 Deep Neural Networks (DNNs) in conjunction with five different source code representations; Abstract Syntax Trees (ASTs), Code Gadgets (CGs), Semantics-based Vulnerability Candidates (SeVCs), Lexed Code Representations (LCRs), and Composite Code Representations (CCRs). Experimental results show that employing different ML strategies in conjunction with the base model structure influences the performance results to a varying degree. However, ML-based techniques suffer from poor performance on class imbalance handling when used in conjunction with source code representations for software vulnerability detection.

Accidents in Nuclear Power Plants (NPPs) can occur for a variety of causes. However, among these, the scale of accidents due to human error can be greater than expected. Accordingly, researches are being actively conducted using artificial intelligence to reduce human error. Most of the research shows high performance based on the numerical data on NPPs, but the expandability of researches using only numerical data is limited. Therefore, in this study, abnormal diagnosis was performed using artificial intelligence based on image data. The methods applied to abnormal diagnosis are the deep neural network, convolution neural network, and convolution recurrent neural network. Consequently, in nuclear power plants, it is expected that the application of more methodologies can be expanded not only in numerical data but also in image-based data.

Edge nodes are crucial for detection against multitudes of cyber attacks on Internet-of-Things endpoints and is set to become part of a multi-billion industry. The resource constraints in this novel network infrastructure tier constricts the deployment of existing Network Intrusion Detection System with Deep Learning models (DLM). We address this issue by developing a novel light, fast and accurate `Edge-Detect' model, which detects Distributed Denial of Service attack on edge nodes using DLM techniques. Our model can work within resource restrictions i.e. low power, memory and processing capabilities, to produce accurate results at a meaningful pace. It is built by creating layers of Long Short-Term Memory or Gated Recurrent Unit based cells, which are known for their excellent representation of sequential data. We designed a practical data science pipeline with Recurring Neural Network to learn from the network packet behavior in order to identify whether it is normal or attack-oriented. The model evaluation is from deployment on actual edge node represented by Raspberry Pi using current cybersecurity dataset (UNSW2015). Our results demonstrate that in comparison to conventional DLM techniques, our model maintains a high testing accuracy of 99% even with lower resource utilization in terms of cpu and memory. In addition, it is nearly 3 times smaller in size than the state-of-art model and yet requires a much lower testing time.

The emergence of the 5G network has boosted the advancements in the field of the internet of things (IoT) and edge/cloud computing. We present a novel architecture to detect fire in indoor and outdoor environments, dubbed as EAC-FD, an abbreviation of edge and cloud-based fire detection. Compared with existing frameworks, ours is lightweight, secure, cost-effective, and reliable. It utilizes a hybrid edge and cloud computing framework with Intel neural compute stick 2 (NCS2) accelerator is for inference in real-time with Raspberry Pi 3B as an edge device. Our fire detection model runs on the edge device while also capable of cloud computing for more robust analysis making it a secure system. We compare different versions of SSD-MobileNet architectures with ours suitable for low-end devices. The fire detection model shows a good balance between computational cost frames per second (FPS) and accuracy.

A smart home provides better living environment by allowing remote Internet access for controlling the home appliances and devices. Security of smart homes is an important application area commonly using Passive Infrared Sensors (PIRs), image capture and analysis but such solutions sometimes fail to detect an event. An unambiguous person detection is important for security applications so that no event is missed and also that there are no false alarms which result in waste of resources. Cloud platforms provide deep learning and IoT services which can be used to implement an automated and failsafe security application. In this paper, we demonstrate reliable person detection for indoor and outdoor scenarios by integrating an application running on an edge device with AWS cloud services. We provide results for identifying a person before authorizing entry, detecting any trespassing within the boundaries, and monitoring movements within the home.

Deep learning has made remarkable achievements in various domains. Active learning, which aims to reduce the budget for training a machine-learning model, is especially useful for the Deep learning tasks with the demand of a large number of labeled samples. Unfortunately, our empirical study finds that many of the active learning heuristics are not effective when applied to Deep learning models in batch settings. To tackle these limitations, we propose a density weighted diversity based query strategy (DWDS), which makes use of the geometry of the samples. Within a limited labeling budget, DWDS enhances model performance by querying labels for the new training samples with the maximum informativeness and representativeness. Furthermore, we propose a beam-search based method to obtain a good approximation to the optimum of such samples. Our experiments show that DWDS outperforms existing algorithms in Deep learning tasks.

Recently, federated learning (FL), as an advanced and practical solution, has been applied to deal with privacy-preserving issues in distributed multi-party federated modeling. However, most existing FL methods focus on the same privacy-preserving budget while ignoring various privacy requirements of participants. In this paper, we for the first time propose an algorithm (PLU-FedOA) to optimize the deep neural network of horizontal FL with personalized local differential privacy. For such considerations, we design two approaches: PLU, which allows clients to upload local updates under differential privacy-preserving of personally selected privacy level, and FedOA, which helps the server aggregates local parameters with optimized weight in mixed privacy-preserving scenarios. Moreover, we theoretically analyze the effect on privacy and optimization of our approaches. Finally, we verify PLU-FedOA on real-world datasets.

Federated learning is a distributed learning technique where machine learning models are trained on client devices in which the local training data resides. The training is coordinated via a central server which is, typically, controlled by the intended owner of the resulting model. By avoiding the need to transport the training data to the central server, federated learning improves privacy and efficiency. But it raises the risk of model theft by clients because the resulting model is available on every client device. Even if the application software used for local training may attempt to prevent direct access to the model, a malicious client may bypass any such restrictions by reverse engineering the application software. Watermarking is a well-known deterrence method against model theft by providing the means for model owners to demonstrate ownership of their models. Several recent deep neural network (DNN) watermarking techniques use backdooring: training the models with additional mislabeled data. Backdooring requires full access to the training data and control of the training process. This is feasible when a single party trains the model in a centralized manner, but not in a federated learning setting where the training process and training data are distributed among several client devices. In this paper, we present WAFFLE, the first approach to watermark DNN models trained using federated learning. It introduces a retraining step at the server after each aggregation of local models into the global model. We show that WAFFLE efficiently embeds a resilient watermark into models incurring only negligible degradation in test accuracy (-0.17%), and does not require access to training data. We also introduce a novel technique to generate the backdoor used as a watermark. It outperforms prior techniques, imposing no communication, and low computational (+3.2%) overhead$^\textrm1$$^\textrm1$\$The research report version of this paper is also available in https://arxiv.org/abs/2008.07298, and the code for reproducing our work can be found at https://github.com/ssg-research/WAFFLE.

Deep learning has been used in a wide range of applications like computer vision, natural language processing and image detection. The advancement in deep learning algorithms in image detection and manipulation has led to the creation of deepfakes, deepfakes use deep learning algorithms to create fake images that are at times very hard to distinguish from real images. With the rising concern around personal privacy and security, Many methods to detect deepfake images have emerged, in this paper the use of deep learning for creating as well as detecting deepfakes is explored, this paper also propose the use of deep learning image enhancement method to improve the quality of deepfakes created.

This paper aims to address the problem of Deepfake Detection using a Two-Stream Capsule Network. First we review methods used to create Deepfake content, as well as methods proposed in the literature to detect such Deepfake content. We then propose a novel architecture to detect Deepfakes, which consists of a two-stream Capsule network running in parallel that takes in both RGB images/frames as well as Error Level Analysis images. Results show that the proposed approach exhibits the detection accuracy of 73.39 % and 57.45 % for the Deepfake Detection Challenge (DFDC) and the Celeb-DF datasets respectively. These results are, however, from a preliminary implementation of the proposed approach. As part of future work, population-based optimization techniques such as Particle Swarm Optimization (PSO) will be used to tune the hyper parameters for better performance.

Deepfake detection techniques have been widely studied to resolve security issues. However, existing techniques mainly focused on RGB channel-based analysis, which still shows incomplete detection accuracy. In this paper, we validate the performance of Gray channel-based deepfake detection. To compare RGB channel-based analysis and Gray channel-based analysis in deepfake detection, we quantitatively measured the performance by using popular CNN models, deepfake datasets, and evaluation indicators. Our experimental results confirm that Gray channel-based deepfake detection outperforms RGB channel-based deepfake detection in terms of accuracy and analysis time.

Video analytics has many applications in traffic control, security monitoring, action/event analysis, etc. With the adoption of deep neural networks, the accuracy of video analytics in video streams has been greatly improved. However, deep neural networks for performing video analytics are compute-intensive. In order to reduce processing time, many systems switch to the lower frame rate or resolution. State-of-the-art switching approaches adjust configurations by profiling video clips on a large configuration space. Multiple configurations are tested periodically and the cheapest one with a desired accuracy is adopted. In this paper, we propose a method that adapts the configuration by analyzing past video analytics results instead of profiling candidate configurations. Our method adopts a lower/higher resolution or frame rate when objects move slow/fast. We train a model that automatically selects the best configuration. We evaluate our method with two real-world video analytics applications: traffic tracking and pose estimation. Compared to the periodic profiling method, our method achieves 3%-12% higher accuracy with the same resource cost and 8-17x faster with comparable accuracy.

Nowadays, developing an intelligent system for segmenting the moving object from the background is essential task for video surveillance applications. Recently, a deep learning segmentation algorithm composed of encoder CNN, a Feature Pooling Module and a decoder CNN called FgSegNET\_S has been proposed. It is capable to train the model using few training examples. FgSegNET\_S is relying only on the spatial information while it is fundamental to include temporal information to distinguish if an object is moving or not. In this paper, an improved version known as (T\_FgSegNET\_S) is proposed by using the subtracted images from the initial background as input. The proposed approach is trained and evaluated using two publicly available infrared datasets: remote scene infrared videos captured by medium-wave infrared (MWIR) sensors and the Grayscale Thermal Foreground Detection (GTFD) dataset. The performance of network is evaluated using precision, recall, and F-measure metrics. The experiments show improved results, especially when compared to other state-of-the-art methods.

The main objective of Human Activity Recognition (HAR) is to detect various activities in video frames. Video surveillance is an import application for various security reasons, therefore it is essential to classify activities as usual and unusual. This paper implements the deep learning model that has the ability to classify and localize the activities detected using a Single Shot Detector (SSD) algorithm with a bounding box, which is explicitly trained to detect usual and unusual activities for security surveillance applications. Further this model can be deployed in public places to improve safety and security of individuals. The SSD model is designed and trained using transfer learning approach. Performance evaluation metrics are visualised using Tensor Board tool. This paper further discusses the challenges in real-time implementation.

Nowadays, video surveillance cameras are widely used in many smart city applications for ensuring road safety. We can use video data from them to solve such tasks as traffic management, driving control, environmental monitoring, etc. Most of these applications are based on object recognition and tracking algorithms. However, the video image quality is not always meet the requirements of such algorithms due to the influence of different external factors. A variety of adverse weather conditions produce noise on the images, which often makes it difficult to detect objects correctly. Lately, deep learning methods show good results in image processing, including denoising tasks. This work is devoted to the study of using these methods for image quality enhancement in difficult weather conditions such as snow, rain, fog. Different deep learning techniques were evaluated in terms of their impact on the quality of object detection/recognition. Finally, the system for automatic image denoising was developed.