Deep Learning Research Papers - Academia.edu (original) (raw)

- by
- •
- Knowledge Creation, Learning theories, Deep Learning, Correlation
- by Tudor Florin Ursuleanu and +1
- •
- Computer Science, Artificial Intelligence, Surgery, Image Analysis

Human activity recognition based on sensor data is a topic with great potential for customized healthcare. Here, an end-to-end deep learning architecture for human activity/transition recognition is developed, achieving an error rate of 0.82%. Various deep learning models are analyzed, and a hyperparameter search is conducted to optimize our chosen model. First, an LSTM architecture is examined, which has the advantage of allowing variable-length input sequences for both training and inference. However, our best architecture (HATRNet) is a deep convolutional neural network with late sensor fusion i.e. separate processing pipelines for subsets of the input channels. We feed in zero-padded time sequences to our network, and achieve accuracy exceeding the state-of-the-art reported in literature--all without the use of hand-extracted features.

- by Akash Levy
- •
- Machine Learning, Artificial Neural Networks, Deep Learning, Personalized Healthcare

With the increasing utilization of the Internet and its provided services, an increase in cyber-attacks to exploit the information occurs. A technology to store and maintain user's information that is mostly used for its simplicity and low-cost services is cloud computing (CC). Also, a new model of computing that is noteworthy today is mobile cloud computing (MCC) that is used to reduce the limitations of mobile devices by allowing them to offload certain computations to the remote cloud. The cloud environment may consist of critical or essential information of an organization; therefore, to prevent this environment from possible attacks a security solution is needed. An intrusion detection system (IDS) is a solution to these security issues. An IDS is a hardware or software device that can examine all inside and outside network activities and recognize doubtful patterns that may demonstrate a network attack and automatically alert the network (or system) administrator. Because of the ability of an IDS to detect known/unknown (inside/outside) attacks, it is an excellent choice for securing cloud computing. Various methods are used in an intrusion detection system to recognize attacks more accurately. Unlike survey papers presented so far, this paper aims to present a comprehensive survey of intrusion detection systems that use computational intelligence (CI) methods in a (mobile) cloud environment. We firstly provide an overview of CC and MCC paradigms and service models, also reviewing security threats in these contexts. Previous literature is critically surveyed, highlighting the advantages and limitations of previous work. Then we define a taxonomy for IDS and classify CI-based techniques into single and hybrid methods. Finally, we highlight open issues and future directions for research on this topic.

3
rd International Conference on Data Mining & Machine Learning (DMML 2022) will act as
a major forum for the presentation of innovative ideas, approaches, developments, and research
projects in the areas of Data Mining and Machine Learning. It will also serve to facilitate the
exchange of information between researchers and industry professionals to discuss the latest issues
and advancement in the area of Big Data and Machine Learning.
Authors are solicited to contribute to the conference by submitting articles that illustrate research
results, projects, surveying works and industrial experiences that describe significant advances in
Data Mining and Machine Learning.

Although artificial intelligence is the product of science-fiction literature, it currently represents a significant branch of computer science dealing with intelligent behavior, machine learning, and machine adaptation. It has become a discipline that attempts to answer real-world problems. Artificial intelligence systems are nowadays widely used in economics and medicine, design or military. The role of archives is changing worldwide. In this grandiose transformation, archives need to be at the forefront of their own future, so that they can steer, guide, and not lose out.
The vast masses of information in archives provide an excellent platform for the exploitation of artificial intelligence. The mass of data can be a great help not only for research but also for policy preparation and in some areas of public administration in the not too distant future.

- by István Hegedűs
- •
- Artificial Intelligence, Archival Studies, Machine Learning, Archives

The proliferation of data together with the increase of computing power in the last decade has triggered a new interest in artificial intelligence methods. Machine learning and in particular deep learning techniques, inspired by the topological structure of neurons network in brains, are omnipresent in the IT discourse, and generated new enthusiasms and fears in our society. These methods have already shown great effectiveness in fields far from architecture and have long been exploited in software that we use every day. Many computing libraries are available for anyone with some programming skills and allow them to "train" a neural network based on several types of data. The world of architecture has not remained external to this phenomenon: many researchers are working on the applications of artificial intelligence to architectural design, a few design software allow exploiting machine learning algorithms, and some large architectural firms have begun to experiment with deep learning methods to put into practice data accumulated over years of profession, with a special interest in environmental sustainability and building performance. If on the one hand, these techniques promise great results, on the other we are still in an exploratory phase. It is then necessary, in our opinion, to understand what the roles of this technology could be within the architectural design process, and with which scopes they can facilitate such a complex profession as that of the architect. On this subject we made ten interviews with as many designers and researchers in the AEC industry, In the article we will report a summary of their testimonies, comparing and commenting on the responses of the designers, with the aim of understanding the potentials of using artificial intelligence methods within the design process, report their perceptions on how artificial intelligence techniques can affect the architect's approach to the project, concluding with some reflections on the critical issues identified during the interviews with the designers.

- by Giuseppe Gallo and +1
- •
- Artificial Intelligence, Architecture, Machine Learning, Design Process (Architecture)

The stock market is known for its extreme complexity and volatility, and people are always
looking for an accurate and effective way to guide stock trading. In addition, LSTM avoids
long-term dependence issues due to its unique storage unit structure, and it helps predict
financial time series. Based on LSTM and an attention mechanism, a wavelet transform is
used to denoise historical stock data, extract and train its features, and establish the prediction
model of a stock price. We have taken Apple Inc. dataset from Yahoo Finance API which
contains the closing stock prices of the company. The accuracy which we achieved is 91%
and the Mean Squared Error (MSE) which we got is 9%.

- by Srikar Goud Ailla
- •
- Artificial Intelligence, Machine Learning, Deep Learning

We examine Dropout through the perspective of interactions: learned effects that combine multiple input variables. Given NNN variables, there are O(N2)O(N^2)O(N2) possible pairwise interactions, O(N3)O(N^3)O(N3) possible 3-way interactions, etc. We show that Dropout implicitly sets a learning rate for interaction effects that decays exponentially with the size of the interaction, corresponding to a regularizer that balances against the hypothesis space which grows exponentially with number of variables in the interaction. This understanding of Dropout has implications for the optimal Dropout rate: higher Dropout rates should be used when we need stronger regularization against spurious high-order interactions. This perspective also issues caution against using Dropout to measure term saliency because Dropout regularizes against terms for high-order interactions. Finally, this view of Dropout as a regularizer of interaction effects provides insight into the varying effectiveness of Dropout for diffe...

- by Ben Lengerich and +1
- •
- Computer Science, Artificial Intelligence, Deep Learning, Deep Neural Networks
- by DR MD FIROJ ALAM
- •
- Computer Science, Artificial Intelligence, Social Media, Benchmarking
- by Selman Kunduraci
- •
- Machine Learning, Mobile Technology, Deep Learning, Mobile Software

Generating a description of an image is called image captioning. Image captioning requires to recognize the important objects, their attributes and their relationships in an image. It also needs to generate syntactically and semantically correct sentences. Deep learning-based techniques are capable of handling the complexities and challenges of image captioning. In this survey paper, we aim to present a comprehensive review of existing deep learning-based image captioning techniques. We discuss the foundation of the techniques to analyze their performances, strengths and limitations. We also discuss the datasets and the evaluation metrics popularly used in deep learning based automatic image captioning.

- by Md. Zakir Hossain
- •
- Mathematics, Computer Science, Artificial Intelligence, Deep Learning
- by Donald Brown
- •
- Transportation Engineering, Safety Engineering, Mathematics, Computer Science

The learning process in humans requires continuous contacts with environmental stimuli, especially during neurodevelopmental growth. These functions are assisted by the coding potential of mirror neurons to serve social interactions. This ability to learn imitating the observed behavior is no longer necessary during adulthood, and control mechanisms prevent automatic mirroring. However, children with Gilles de la Tourette syndrome could encounter coding errors at the level of the mirror neurons system as these cortical regions are themselves the ones affected in the syndrome. Combined with impulsivity, the resulting sign would be a manifest echopraxia that persists throughout adulthood, averting these individuals from the appraisal of a spot-on motor control.

- by Matteo Briguglio
- •
- Deep Learning, Mirror Neuron, Echopraxia, Motor Tic

Background In recent years, deep learning has gained remarkable attention in medical image analysis due to its capacity to provide results comparable to specialists and, in some cases, surpass them. Despite the emergence of deep learning research on gastric tissues diseases, few intensive reviews are addressing this topic. Method We performed a systematic review related to applications of deep learning in gastric tissue disease analysis by digital histology, endoscopy and radiology images. Conclusions This review highlighted the high potential and shortcomings in deep learning research studies applied to gastric cancer, ulcer, gastritis and non-malignant diseases. Our results demonstrate the effectiveness of gastric tissue analysis by deep learning applications. Moreover, we also identified gaps of evaluation metrics, and image collection availability, therefore, impacting experimental reproducibility.

- by Wanderson Gonçalves e Gonçalves and +1
- •
- Artificial Intelligence, Machine Learning, Gastroenterology, Computational Intelligence

The purpose of this article is to estimate the purchasing and sale opportunities of houses on the market by Machine learning techniques. For financial stability, the housing sector is quite critical. People seeking to purchase a new house appear to be more cautious in their expectations and sales tactics analyzing historical industry patterns and pricing levels, as well as potential changes. The index of our method consists of the price of the house and its efficiency metrics, such as the amount of renovation, the distance from the city center, the construction programs, the height of the property, the floor and the location of the apartment in the home, and there are several other criteria. Service includes a database that recognizes the preferences of its clients and then integrates machine learning software. The program will enable consumers invest in real estate without approaching brokers. It therefore reduces the uncertainties inherent with the deal. The program has a login ID and a pin. At the same time, when the user searches for an attribute, the value of the original attribute and the value of the predicted attribute are displayed.

- by Anirban Chakraborty
- •
- Regression Models, Artificial Neural Networks, Deep Learning

In this review, the application of in-depth learning for medical diagnosis will be corrected. A thorough analysis of various scientific articles in the domain of deep neural network applications in the medical field has been implemented. Has received more than 300 research articles and after several steps of selection, 46 articles have been presented in more detail The research found that the neural network (CNN) is the most prevalent agent when talking about deep learning and medical image analysis. In addition, from the findings of this article, it can be observed that the application of widespread learning technology is widespread. But most of the applications that focus on bioinformatics, medical diagnostics and other similar fields. In this work, we examine the strength of the deep learning method for pathological examination in chest radiography. Convolutional neural networks (CNN) The method of deep architectural classification is popular due to the ability to learn to represent medium and high level images. We explore CNN's ability to identify different types of diseases in chest X-ray images. Moreover, because of the very large training sets that are not available in the medical domain, we therefore explore the possibility of using deep learning methods based on non-medical learning. We tested our algorithm on 93 datasets. We use CNN that is trained with ImageNet, which is a well-known non-animated large image database. The best performance is due to the use of features pulled from CNN and low-level features.

- by Vrushali Dhanokar
- •
- Computer Science, Image Processing, Machine Learning, Data Mining

Background: Brain-Computer Interface (BCI) is becoming more reliable, thanks to the advantages of Artificial Intelligence (AI). Recently, hybrid Deep Learning (hDL), which combines different DL algorithms, has gained momentum over the past five years. In this work, we proposed a review on hDL-based BCI starting from the seminal studies in 2015. Objectives: We have reviewed 47 papers that apply hDL to the BCI system published between 2015 and 2020 extracting trends and highlighting relevant aspects to the topic. Methods: We have queried four scientific search engines (Google Scholar, PubMed, IEEE Xplore and Elsevier Science Direct) and different data items were extracted from each paper such as the database used, kind of application, online/offline training, tasks used for the BCI, pre-processing methodology adopted, type of normalization used, which kind of features were extracted, type of DL architecture used, number of layers implemented and which optimization approach were used a...

- by Nibras Abo Alzahab
- •
- Non-Invasive BCI, Medicine, BCI, EEG Signal Processing

With the rapid development of numerous wireless network technologies and the growing number of wireless devices in use around the world, gaining access to the radio frequency spectrum has become a challenge that must be solved as soon as possible. The ever-increasing wireless traffic and shortage of accessible spectrum necessitate smart spectrum management. Machine learning (ML) is gaining popularity, and its capacity to spot patterns and aid decision-making has found applications in a variety of disciplines. Machine learning approaches have been applied to wireless networking difficulties, such as spectrum efficiency, and have showed superior performance compared to traditional methods. Spectrum sensing enables dynamic spectrum sharing, which improves spectrum efficiency by allowing coexistence of wireless technologies within the same frequency range. This involves the accurate detection and identification of multiple wireless signals sent in the same radio spectrum range. The current state of machine learning algorithms for identifying and classifying radio signals depending on their access technologies, such as Wi-Fi and LTE, is examined in this work. Classifying the RF signals based on their wireless network technologies as opposed to their modulation schemes, especially using machine learning, is an emerging area of study and is becoming a popular research topic. This survey will assist readers to become familiar with the current literature and enable further research in this domain.

A computer views all kinds of visual media as an array of numerical values. As a consequence of this approach, they require image processing algorithms to inspect contents of images. This project compares 3 major image processing algorithms: Single Shot Detection (SSD), Faster Region based Convolutional Neural Networks (Faster R-CNN), and You Only Look Once (YOLO) to find the fastest and most efficient of three. In this comparative analysis, using the Microsoft COCO (Common Object in Context) dataset, the performance of these three algorithms is evaluated and their strengths and limitations are analysed based on parameters such as accuracy, precision and F1 score. From the results of the analysis, it can be concluded that the suitability of any of the algorithms over the other two is dictated to a great extent by the use cases they are applied in. In an identical testing environment, YOLO-v3 outperforms SSD and Faster R-CNN, making it the best of the three algorithms.

- by Ved Kulkarni
- •
- Computer Science, Artificial Intelligence, Deep Learning, Big Data

In computer networks, transmitted traffic between origin-destination nodes has been considered a crucial factor in traffic engineering applications. To date, measuring the traffic directly in high load networks is difficult due to high computational costs. On the other hand, accurate estimation of network traffic by means of link load measurements and routing information is currently a challenging problem. In this paper, we propose a new approach to estimate end-to-end traffic, inspired by graph embedding. In the proposed approach, we consider a computer network as a time-varying graph. Our model provides explicit routing information to a convolutional neural network (ConvNet) estimator via link load measurements and network topological structure. When explicit routing information is provided, the learning model is only expected to embed the relations between link loads into a traffic estimation vector, instead of figuring out the routing paths. The experimental results showed that the proposed approach outperforms similar estimators in terms of lower estimation error and better tracking the fluctuations.

- by Ali Zamani and +1
- •
- Computer Science, Machine Learning, Deep Learning, Traffic Matrix

Recent advancements in activity recognition from sports videos have attracted wide scientific interest of the Computer Vision community. However, the activity recognition problem from cricket video sequences is largely under-represented in the literature. This paper aims to devise a convolutional neural network (CNN) based model for sports activity recognition. The model is trained on the pre-trained VGG16, VGG19, ResNet50, and Inception V3 Models and tested on the clustered cricket videos frames extracted from the data set especially prepared for this research. The clustering of the frames is done by using K-Mean clustering algorithm. K-Fold cross validation is done which gave an accuracy of 99% on clustered data and 91% on un-clustered data. The accuracy and time complexity of the proposed method is better as compared to the state of the art methods used for activity recognition from videos.

- by Malik Tabish
- •
- Image Processing, Machine Learning, Video Processing, Deep Learning

Much recent research has been devoted to learning algorithms for deep architectures such as Deep Belief Networks and stacks of auto-encoder variants, with impressive results obtained in several areas, mostly on vision and language data sets. The best results obtained on supervised learning tasks involve an unsupervised learning component, usually in an unsupervised pre-training phase. Even though these new algorithms

- by Yoshua Bengio
- •
- Machine Learning, Non-convex optimization, Learning problems, Deep Learning

This work evaluates deep learning-based myocardial infarction (MI) quantification using Segment cardiovascular magnetic resonance (CMR) software. Segment CMR software incorporates the expectation-maximization, weighted intensity, a priori information (EWA) algorithm used to generate the infarct scar volume, infarct scar percentage, and microvascular obstruction percentage. Here, Segment CMR software segmentation algorithm is updated with semantic segmentation with U-net to achieve and evaluate fully automated or deep learning-based MI quantification. The direct observation of graphs and the number of infarcted and contoured myocardium are two options used to estimate the relationship between deep learning-based MI quantification and medical exper-based results.

- by Olivier Rukundo, Ph.D.
- •
- Artificial Intelligence, Medical Image Processing, MRI, Deep Learning

One of the most famous cultural heritages in Indonesia is batik. Batik is a specially made drawing cloth by writing Malam (wax) on the cloth, then processed in a certain way. The diversity of motifs both in Indonesia and the allied countries raises new research topics in the field of information technology, both for conservation, storage, publication and the creation of new batik motifs. In computer science research area, studies about Batik pattern have been done by researchers and some algorithms have been successfully applied in Batik pattern recognition. This study was focused on Batik motif recognition using texture fusion feature which is Gabor, Log-Gabor, and GLCM; and using PCA feature reduction to improve the classification accuracy and reduce the computational time. To improve the accuracy, we proposed a Deep Neural Network model to recognise batik pattern and used batch normalisation as a regularises to generalise the model and to reduce time complexity. From the experiments, the feature extraction, selection, and reduction gave better accuracy than the raw dataset. The feature selection and reduction also reduce time complexity. The DNN+BN significantly improve the accuracy of the classification model from 65.36% to 83.15%. BN as a regularization has successfully made the model more general, hence improve the accuracy of the model. The parameters tuning also improved accuracy from 83.15% to 85.57%.

Setup Python on your system in 5 minutes.

- by Mohamed Loey
- •
- Deep Learning

This paper presents a deep learning approach for age estimation of human beings using their facial images. The different racial groups based on skin colour have been incorporated in the annotations of the images in the dataset, while ensuring an adequate distribution of subjects across the racial groups so as to achieve an accurate Automatic Facial Age Estimation (AFAE). The principle of transfer learning is applied to the ResNet50 Convolutional Neural Network (CNN) initially pretrained for the task of object classification and finetuning it’s hyperparameters to propose an AFAE system that can be used to automate ages of humans across multiple racial groups. The mean absolute error of 4.25 years is obtained at the end of the research which proved the effectiveness and superiority of the proposed method.

- by Shreshth Tuli
- •
- Computer Science, Software Engineering, Climate Change, Machine Learning

Modeling surface water quality using soft computing techniques is essential for the effective management of scarce water resources and environmental protection. The development of accurate predictive models with significant input parameters and inconsistent datasets is still a challenge. Therefore, further research is needed to improve the performance of the predictive models. This study presents a methodology for dataset pre-processing and input optimization for reducing the modeling complexity. The objective of this study was achieved by employing a two-sided detection approach for outlier removal and an exhaustive search method for selecting essential modeling inputs. Thereafter, the adaptive neuro-fuzzy inference system (ANFIS) was applied for modeling electrical conductivity (EC) and total dissolved solids (TDS) in the upper Indus River. A larger dataset of a 30-year historical period, measured monthly, was utilized in the modeling process. The prediction capacity of the develo...

- by Taher Abunama
- •
- Computer Science, Artificial Intelligence, Machine Learning, Hydrology

During the last few years, spoken language technologies have known a big improvement thanks to Deep Learning. However Deep Learning-based algorithms require amounts of data that are often difficult and costly to gather. Particularly, modeling the variability in speech of different speakers, different styles or different emotions with few data remains challenging. In this paper, we investigate how to leverage fine-tuning on a pre-trained Deep Learning-based TTS model to synthesize speech with a small dataset of another speaker. Then we investigate the possibility to adapt this model to have emotional TTS by fine-tuning the neutral TTS model with a small emotional dataset.

- by Kevin El Haddad
- •
- Emotion, Artificial Intelligence, Human Computer Interaction, Machine Learning

Visual speech information plays an important role in lip-reading under noisy conditions or for listeners with a hearing impairment. Correct utterances to read Quran for beginners, there are rules of utterances to learn Quran and we need a software system to tell us if we utter correctly. For that, we built lip-reading model, the model localizes the lips efficiently.
We present in this study a classification model for some al-tajweed rules as we depended on Machine Learning - Cascade Object Detector (Viola-Jones Algorithm), HOG features, a multiclass SVM classifier and Aggregate Channel Features (ACF) object detector for features extraction. We uses Matlab to train a classifiers using a pre-trained convolutional neural network (CNN) for classifying images from the video stream of four different Rules of Holy Quran Allah Elevating (mufakhum), Allah Lowering (moureqeq), sunny لام and moonyلام . CNN acquires multiple convolutional filters, used to extract visual features essential for recognizing phoneme. CNNs produce

- by safaa omer
- •
- Artificial Intelligence, Image Processing, Machine Learning, Deep Learning

It's a fairly simple task for humans to determine the gender of an individual using certain facial features, although it is difficult for machines to perform an equivalent task. Within the past decade, unimaginable steps have been taken to automatically predict the gender from a face image. The human face has certain distinctive features such as eyes, nose, lips, etc., which can be analyzed to classify humans into two basic genders: Male and Female. This project aims at achieving a similar goal of detecting gender from face images. The basic tool used in the project is Convolutional Neural Network (CNN) along with the use of the Programming language Python. In recent years, face detection has achieved considerable attention from researchers in biometrics, pattern recognition, and computer vision groups. There are countless security and forensic applications requiring the use of face recognition technologies which have motivated us to explore this area and start with this ...

- by Devjyoti Saha
- •
- Computer Vision, Deep Learning, Convolutional Neural Network [CNN]

Machine Learning and Applications: An International Journal (MLAIJ) is a quarterly open access peer-reviewed journal that publishes articles which contribute new results in all areas of the machine learning. The journal is devoted to the publication of high quality papers on theoretical and practical aspects of machine learning and applications. The goal of this journal is to bring together researchers and practitioners from academia and industry to focus on machine learning advancements, and establishing new collaborations in these areas. Original research papers, state-of-the-art reviews are invited for publication in all areas of machine learning.

Recently, the field of deep learning has received increased attention due to its high accuracy. A common deep learning technique is Convolutional Neural Networks (CNN), which is as a construction of trainable multi-stages using multiple phases. In this paper, we use a type of CNN called AlexNet to classify human retinal images into 'normal' or 'have been treated using photocoagulation laser treatments' classes. Indeed, this classification technique will help experts to examine any case and make the examination process faster and more efficient. The study was conducted through several experiments using 730 images of human retina that were either treated by laser or not treated. An average accuracy rate of more than 97% was obtained. Additionally, possible improvements and destiny traits are suggested to summarize this study.

- by Mustafa Abuzaraida
- •
- Image Processing, Machine Learning, Deep Learning

Knowing rainforest environments is rendered challenging by distance, vegetation intensity, and coverage; however, knowing the complexity and sustainability of these landscapes is important for ecologists and conservationists. The airborne light detection and ranging (LiDAR) system has made dramatic improvements to forest data collection and management especially on the forest inventory aspect. LiDAR can reliably calculate tree-level characteristics such as crown scale and tree height as well as derived measures such as breast height diameter (DBH). To do this, an exact tree extraction method is needed inside LiDAR data. Within LiDAR data, tree extraction often starts by locating the treetops via local maxima (LM). Wide-ranging efforts have been developed to extract individual trees from LiDAR data by starting to localize treetops through LM within LiDAR data. Throughout this research, a demonstration of a new tree extraction framework inside LiDAR Point Cloud by incorporating a new tree extraction method using the bounding-box coordinates provided by deep learning-based object detection. Tree extraction inside the LiDAR point cloud using the bounding-box coordinates was successful and feasible.

- by Alvin Sarraga Alon
- •
- Artificial Intelligence, Remote Sensing, Machine Learning, Embedded Systems

Victims of road traffic accidents face severe health problems on-site or after the event when they arrive at hospital lately in their emergency cycle. Road traffic accident has negative effect on the physical, social and emotional security of human lives which often lead to mortality, illness, pain, grief and even disability. This paper proposes a scheme that reduces the severity of road traffic accidents given its inevitable occurrence. The rational is to search for nearest hospitals to the accident location using Dijkstra algorithm and Fuzzy logic to recommend suitable hospitals out of list of nearest hospitals to timely attend to the emergency situation considering factors such as distance, severity of the accident, available facilities in the hospitals and other factors. The obtained results showed the practicability of the system to recommendation of quick solution to accident emergencies.

Thermal imagery is a substitute of visible imagery for face detection due to its property of illumination invariance with the variation of facial appearances. This paper presents an effective method for human face detection in thermal imaging. The concept of histogram plot has been used in the feature extraction process and later in face detection. Techniques like thresholding, object boundary analysis, morphological operation etc. have been performed on the images to ease the process of detection. In order to enhance the performance of the algorithm and to reduce the computation time, parallelism has been achieved using Message Passing Interface (MPI) model. Overall, the proposed algorithm showed a higher level of accuracy and less complexity time of 0.11 seconds in the parallel environment as compared to 0.20 seconds in a serial environment.

- by Chirag Kyal
- •
- Parallel Computing, Computer Vision, Image Processing, Machine Learning

In today's world women safety is one of the most important issues to be addressed in our country. When a women needs urgent help at the time of harassment or molestation, proper reachability is not present for them. Apart from being aware about the significance of women's safety, it is essential that they are provided with protection during those crucial times. The earlier existing system are helpful in detecting the women's location after the crime has been committed. In this project we will be using the women's handbag in which we will be fixing camera lenses and which will be carried anywhere they go. Whenever she comes in contact with any person outside, an image of that person is taken and the activities of the person can be monitored continuously. If the person behaves normally the image can be of no use and can be deleted. But if the activities of the person varies resulting in any harmful action then our system will detect it and process the captured image and it will send to the police and family members with GPS location tracked from IP address. Thus our project helps in saving the life of a women and safeguarding her in the present situation.

- by IJCSMC Journal
- •
- Mathematics, Information Technology, Technology, Neural Networks

Obtaining an effective reward signal for dialogue management is a non trivial problem. Real user feedback is inconsistent and often even absent. This thesis investigates the use of intrinsic rewards for a reinforcement learning based dialogue manager in order to improve policy learning in an environment with sparse rewards and to move away from inefficient random ε-greedy exploration. In addition to rewards given by a user simulator for successful dialogues, intrinsic curiosity rewards are given in the form of belief-state prediction errors generated by an intrinsic curiosity module within the dialogue manager. We investigate two main settings for this method: (1) predicting the raw next belief-state, and (2) predicting belief-states in a learned feature space. In order to meet the right difficulty level for the system to be able to learn a feature space, the model is pre-trained on a small pool of dialogue transitions. For both settings, results comparable to and better than simple ε-greedy exploration are achieved. (1) is able to learn faster, but (2) achieves higher final results and has more potential for improvements and to be successful in larger state-action spaces, where feature encodings and generalization are beneficial.

- by Michael Chinkers
- •
- Robotics, Artificial Intelligence, Reinforcement Learning, Dialogue

The emerging field of quantum machine learning has the potential to substantially aid in the problems and scope of artificial intelligence. This is only enhanced by recent successes in the field of classical machine learning. In this work we propose an approach for the systematic treatment of machine learning, from the perspective of quantum information. Our approach is general and covers all three main branches of machine learning: supervised, unsupervised, and reinforcement learning. While quantum improvements in supervised and unsupervised learning have been reported, reinforcement learning has received much less attention. Within our approach, we tackle the problem of quantum enhancements in reinforcement learning as well, and propose a systematic scheme for providing improvements. As an example, we show that quadratic improvements in learning efficiency, and exponential improvements in performance over limited time periods, can be obtained for a broad class of learning problems.

- by Mohamed Loey
- •
- Deep Learning

In recent times the field of computer vision and deep learning has seen many advancements that have helped in efficient and accurate face and object detection, resulting in many security and surveillance applications. Furthermore, with the increase in localized cameras monitoring every human action, every vehicle tracked it only seemed logical to use these camera's video feeds more efficiently and smartly. Present security systems include manual surveillance or a single smart camera setup. In this paper, we propose a smart multi-camera system using a Resnet-34 model for face recognition and Tesseract-based Optical Character Recognition for vehicle number plate recognition. The individuals and vehicles captured in the multiple cameras are traced out on a map which will graphically show when and where they were detected. This security and surveillance system will be beneficial in large organizations such as offices, banks, shopping malls, residential areas, etc.

- by Chaitanya Sonavane and +2
- •
- Machine Learning, Data Science, Deep Learning, Smart surveillance

In recent times, organisations in a variety of businesses, such as healthcare, education, and others, have been using the Internet of Things (IoT) to produce more competent and improved services. The widespread use of IoT devices makes our lives easier. On the other hand, the IoT devices that we use suffer vulnerabilities that may impact our lives. These unsafe devices accelerate and ease cybersecurity attacks, specifically when using a botnet. Moreover, restrictions on IoT device resources, such as limitations in power consumption and the central processing unit and memory, intensify this issue because they limit the security techniques that can be used to protect IoT devices. Fortunately, botnets go through different stages before they can start attacks, and they can be detected in the early stage. This research paper proposes a framework focusing on detecting an IoT botnet in the early stage. An empirical experiment was conducted to investigate the behaviour of the early stage of the botnet, and then a baseline machine learning model was implemented for early detection. Furthermore, the authors developed an effective detection method, namely, Cross CNN_LSTM, to detect the IoT botnet based on using fusion deep learning models of a convolutional neural network (CNN) and long short-term memory (LSTM). According to the conducted experiments, the results show that the suggested model is accurate and outperforms some of the state-of-the-art methods, and it achieves 99.7 accuracy. Finally, the authors developed a kill chain model to prevent IoT botnet attacks in the early stage.

- by Daniyal Alghazzawi and +1
- •
- Deep Learning, Botnets

The importance of the detection of aggressiveness in social media is due to real effects of violence provoked by negative behavior online. Indeed, this kind of legal cases are increasing in the last years. For this reason, the necessity of controlling user-generated contents has become one of the priorities for many Internet companies, although current methodologies are far from solving this problem. Therefore, in this work we propose an innovative approach that combines deep learning framework with linguistic features specific for this issue. This approach has been evaluated and compared with other ones in the framework of the MEX-A3T shared task at IberEval on aggressiveness analysis in Spanish Mexican tweets. In spite of our novel approach, we obtained low results.

- by Simona Frenda
- •
- Computer Science, Hate Speech, Computational Linguistics & NLP, Deep Learning