Abstract: This paper aims to provide an interpretation of artificial neural networks (ANNs) and explore some of its implications. The interpretation views ANNs as a memory which encodes instances of experience. An experiment explores the behavior of encoding and retrieval of instances from memory. A localised representation ANN is created that allows control over encoding and retrieved memory sample size and is experimented with using the MNIST digits dataset. The relationship between input familiarity, conflict within retrieved samples, and error rates is described and demonstrated to be an effective driver for memory encoding. Results indicate that selective encoding and retrieval samples that allow detection of memory conflicts produce optimal performance, and that error rates are normally distributed with input familiarity and conflict. By using input familiarity and sample consistency to guide memory encoding, the number of encoding trials on the dataset were reduced to 18.33% of the training data while maintaining good recognition performance on the test data.
Abstract: Acoustic sensors, if embedded in smart street lights, can help in capturing the activities (car honking, sirens, events, traffic, etc.) in cities. Needless to say, the acoustic data from such scenarios are complex due to multiple audio streams originating from different events, and when decomposed to independent signals, the amount of retrieved data volume is small in quantity which is inadequate to train deep neural networks. So, in this paper, we address the two challenges: a) separating the mixed signals, and b) developing an efficient acoustic classifier under data paucity. So, to address these challenges, we propose an architecture with supervised deep learning, where the initial captured mixed acoustics data are analyzed with Fast Fourier Transformation (FFT), followed by filtering the noise from the signal, and then decomposed to independent signals by fast independent component analysis (Fast ICA). To address the challenge of data paucity, we propose a multi feature-based deep neural network with high performance that is reflected in our experiments when compared to the conventional convolutional neural network (CNN) and multi-layer perceptron (MLP).
Abstract: Quality of life, environmental impact, congestion of mobility means, and infrastructures remain significant challenges for urban mobility. Solutions like car sharing, spatial redesign, eCommerce, and autonomous vehicles will likely increase the unit veh-km and the density of cars in urban traffic, thus reducing congestion. However, the impact of such solutions is not clear for researchers. Congestion arises from growing populations that must travel greater distances to arrive at similar locations (e.g., workplaces, schools) during the same time frame (e.g., rush hours). This paper first reviews the research and application cases of urban congestion methods through recent years. Rethinking the question of time, it then investigates people’s willingness and flexibility to adapt their arrival and departure times from workplaces. We use neural networks and methods of supervised learning to apply a methodology for predicting peoples’ intentions from their responses in a questionnaire. We created and distributed a questionnaire to more than 50 companies in the Paris suburb. Obtained results illustrate that our methodology can predict peoples’ intentions to reschedule their activities (work, study, commerce, etc.).
Abstract: The Earth's surface is constantly changing through forces of nature and human activities. Reliable, accurate, and timely change detection is critical to environmental monitoring, resource management, and planning activities. Recently, interest in deep learning algorithms, especially convolutional neural networks, has increased in the field of image change detection due to their powerful ability to extract multi-level image features automatically. However, these networks are prone to drawbacks that limit their applications, which reside in their inability to capture spatial relationships between image instances, as this necessitates a large amount of training data. As an alternative, Capsule Network has been proposed to overcome these shortcomings. Although its effectiveness in remote sensing image analysis has been experimentally verified, its application in change detection tasks remains very sparse. Motivated by its greater robustness towards improved hierarchical object representation, this study aims to apply a capsule network for PolSAR image Change Detection. The experimental results demonstrate that the proposed change detection method can yield a significantly higher detection rate compared to methods based on convolutional neural networks.
Abstract: Recent advances in cognitive neuroscience have allowed a step forward in perceiving the processes involved in learning from the point of view of acquiring new information or the modification of existing mental content. The evoked potentials technique reveals how basic brain processes interact to achieve adequate and flexible behaviours. The objective of this work, using evoked potentials, is to study if it is possible to distinguish if a patient suffers a specific type of learning disorder to decide the possible therapies to follow. The methodology used in this work is to analyze the dynamics of different brain areas during a cognitive activity to find the relationships between the other areas analyzed to understand the functioning of neural networks better. Also, the latest advances in neuroscience have revealed the exis-tence of different brain activity in the learning process that can be highlighted through the use of non-invasive, innocuous, low-cost and easy-access techniques such as, among others, the evoked potentials that can help to detect early possible neurodevelopmental difficulties for their subsequent assessment and therapy. From the study of the amplitudes and latencies of the evoked potentials, it is possible to detect brain alterations in the learning process, specifically in dyscalculia, to achieve specific corrective measures for the application of personalized psycho-pedagogical plans that allow obtaining an optimal integral development of the affected people.
Abstract: Depth estimation using stereo images is a challenging problem in computer vision. Many different studies have been carried out to solve this problem. With advancing machine learning, tackling this problem is often done with neural network-based solutions. The images used in these studies are mostly in the visible spectrum. However, the need to use the Infrared (IR) spectrum for depth estimation has emerged because it gives better results than visible spectra in some conditions. At this point, we recommend using thermal-thermal (IR) image pairs for depth estimation. In this study, we used two well-known networks (PSMNet, FADNet) with minor modifications to demonstrate the viability of this idea.
Abstract: One of the most significant issues as attended a lot in recent years is that of recognizing the sentiments and emotions in social media texts. The analysis of sentiments and emotions is intended to recognize the conceptual information such as the opinions, feelings, attitudes and emotions of people towards the products, services, organizations, people, topics, events and features in the written text. These indicate the greatness of the problem space. In the real world, businesses and organizations are always looking for tools to gather ideas, emotions, and directions of people about their products, services, or events related to their own. This article uses the Twitter social network, one of the most popular social networks with about 420 million active users, to extract data. Using this social network, users can share their information and opinions about personal issues, policies, products, events, etc. It can be used with appropriate classification of emotional states due to the availability of its data. In this study, supervised learning and deep neural network algorithms are used to classify the emotional states of Twitter users. The use of deep learning methods to increase the learning capacity of the model is an advantage due to the large amount of available data. Tweets collected on various topics are classified into four classes using a combination of two Bidirectional Long Short Term Memory network and a Convolutional network. The results obtained from this study with an average accuracy of 93%, show good results extracted from the proposed framework and improved accuracy compared to previous work.
Abstract: Identification of plant diseases has been performed using machine learning and deep learning models on the datasets containing images of healthy and diseased plant leaves. The current study carries out an evaluation of some of the deep learning models based on convolutional neural network architectures for identification of plant diseases. For this purpose, the publicly available New Plant Diseases Dataset, an augmented version of PlantVillage dataset, available on Kaggle platform, containing 87,900 images has been used. The dataset contained images of 26 diseases of 14 different plants and images of 12 healthy plants. The CNN models selected for the study presented in this paper are AlexNet, ZFNet, VGGNet (four models), GoogLeNet, and ResNet (three models). The selected models are trained using PyTorch, an open-source machine learning library, on Google Colaboratory. A comparative study has been carried out to analyze the high degree of accuracy achieved using these models. The highest test accuracy and F1-score of 99.59% and 0.996, respectively, were achieved by using GoogLeNet with Mini-batch momentum based gradient descent learning algorithm.
Abstract: This paper proposes a neural network approach for assessing flood hazard mapping. The core of the model is a machine learning component fed by frequency ratios, namely statistical correlations between flood event occurrences and a selected number of topographic properties. The classification capability was compared with the flood hazard mapping River Basin Plans (Piani Assetto Idrogeologico, acronimed as PAI) designed by the Italian Institute for Environmental Research and Defence, ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale), encoding four different increasing flood hazard levels. The study area of Piemonte, an Italian region, has been considered without loss of generality. The frequency ratios may be used as a standalone block to model the flood hazard mapping. Nevertheless, the mixture with a neural network improves the classification power of several percentage points, and may be proposed as a basic tool to model the flood hazard map in a wider scope.
Abstract: Malignant Melanoma, known simply as Melanoma, is a type of skin cancer that appears as a mole on the skin. It is critical to detect this cancer at an early stage because it can spread across the body and may lead to the patient death. When detected early, Melanoma is curable. In this paper we propose a deep learning model (Convolutional Neural Networks) in order to automatically classify skin lesion images as Malignant or Benign. Images underwent certain pre-processing steps to diminish the effect of the normal skin region on the model. The result of the proposed model showed a significant improvement over previous work, achieving an accuracy of 97%.
Abstract: Swimming with the tide of deep learning, the field of music information retrieval (MIR) experiences parallel development and a sheer variety of feature-learning models has been applied to music classification and tagging tasks. Among those learning techniques, the deep convolutional neural networks (CNNs) have been widespreadly used with better performance than the traditional approach especially in music genre classification and prediction. However, regarding the music recommendation, there is a large semantic gap between the corresponding audio genres and the various aspects of a song that influence user preference. In our study, aiming to bridge the gap, we strive to construct an automatic music aesthetic annotation model with MIDI format for better comparison and measurement of the similarity between music pieces in the way of harmonic analysis. We use the matrix of qualification converted from MIDI files as input to train two different classifiers, support vector machine (SVM) and Decision Tree (DT). Experimental results in performance of a tag prediction task have shown that both learning algorithms are capable of extracting high-level properties in an end-to end manner from music information. The proposed model is helpful to learn the audience taste and then the resulting recommendations are likely to appeal to a niche consumer.
Abstract: Analysis of the human microbiome using metagenomic
sequencing data has demonstrated high ability in discriminating
various human diseases. Raw metagenomic sequencing data require
multiple complex and computationally heavy bioinformatics steps
prior to data analysis. Such data contain millions of short sequences
read from the fragmented DNA sequences and stored as fastq files.
Conventional processing pipelines consist in multiple steps including
quality control, filtering, alignment of sequences against genomic
catalogs (genes, species, taxonomic levels, functional pathways,
etc.). These pipelines are complex to use, time consuming and
rely on a large number of parameters that often provide variability
and impact the estimation of the microbiome elements. Training
Deep Neural Networks directly from raw sequencing data is a
promising approach to bypass some of the challenges associated with
mainstream bioinformatics pipelines. Most of these methods use the
concept of word and sentence embeddings that create a meaningful
and numerical representation of DNA sequences, while extracting
features and reducing the dimensionality of the data. In this paper
we present an end-to-end approach that classifies patients into disease
groups directly from raw metagenomic reads: metagenome2vec. This
approach is composed of four steps (i) generating a vocabulary of
k-mers and learning their numerical embeddings; (ii) learning DNA
sequence (read) embeddings; (iii) identifying the genome from which
the sequence is most likely to come and (iv) training a multiple
instance learning classifier which predicts the phenotype based on
the vector representation of the raw data. An attention mechanism
is applied in the network so that the model can be interpreted,
assigning a weight to the influence of the prediction for each genome.
Using two public real-life data-sets as well a simulated one, we
demonstrated that this original approach reaches high performance,
comparable with the state-of-the-art methods applied directly on
processed data though mainstream bioinformatics workflows. These
results are encouraging for this proof of concept work. We believe
that with further dedication, the DNN models have the potential to
surpass mainstream bioinformatics workflows in disease classification
tasks.
Abstract: As biometric systems become widely deployed, the security of identification systems can be easily attacked by various spoof materials. This paper contributes to finding a reliable and practical anti-spoofing method using Convolutional Neural Networks (CNNs) based on the types of loss functions and optimizers. The types of CNNs used in this paper include AlexNet, VGGNet, and ResNet. By using various loss functions including Cross-Entropy, Center Loss, Cosine Proximity, and Hinge Loss, and various loss optimizers which include Adam, SGD, RMSProp, Adadelta, Adagrad, and Nadam, we obtained significant performance changes. We realize that choosing the correct loss function for each model is crucial since different loss functions lead to different errors on the same evaluation. By using a subset of the Livdet 2017 database, we validate our approach to compare the generalization power. It is important to note that we use a subset of LiveDet and the database is the same across all training and testing for each model. This way, we can compare the performance, in terms of generalization, for the unseen data across all different models. The best CNN (AlexNet) with the appropriate loss function and optimizers result in more than 3% of performance gain over the other CNN models with the default loss function and optimizer. In addition to the highest generalization performance, this paper also contains the models with high accuracy associated with parameters and mean average error rates to find the model that consumes the least memory and computation time for training and testing. Although AlexNet has less complexity over other CNN models, it is proven to be very efficient. For practical anti-spoofing systems, the deployed version should use a small amount of memory and should run very fast with high anti-spoofing performance. For our deployed version on smartphones, additional processing steps, such as quantization and pruning algorithms, have been applied in our final model.
Abstract: Predicting the actual cost and duration in construction projects concern a continuous and existing problem for the construction sector. This paper addresses this problem with modern methods and data available from past public construction projects. 39 bridge projects, constructed in Greece, with a similar type of available data were examined. Considering each project’s attributes with the actual cost and the actual duration, correlation analysis is performed and the most appropriate predictive project variables are defined. Additionally, the most efficient subgroup of variables is selected with the use of the WEKA application, through its attribute selection function. The selected variables are used as input neurons for neural network models through correlation analysis. For constructing neural network models, the application FANN Tool is used. The optimum neural network model, for predicting the actual cost, produced a mean squared error with a value of 3.84886e-05 and it was based on the budgeted cost and the quantity of deck concrete. The optimum neural network model, for predicting the actual duration, produced a mean squared error with a value of 5.89463e-05 and it also was based on the budgeted cost and the amount of deck concrete.
Abstract: We as humans use words with accompanying visual and facial cues to communicate effectively. Classifying facial emotion using computer vision methodologies has been an active research area in the computer vision field. In this paper, we propose a simple method for facial expression recognition that enhances accuracy. We tested our method on the FER-2013 dataset that contains static images. Instead of using Histogram equalization to preprocess the dataset, we used Unsharp Mask to emphasize texture and details and sharpened the edges. We also used ImageDataGenerator from Keras library for data augmentation. Then we used Convolutional Neural Networks (CNN) model to classify the images into 7 different facial expressions, yielding an accuracy of 69.46% on the test set. Our results show that using image preprocessing such as the sharpening technique for a CNN model can improve the performance, even when the CNN model is relatively simple.
Abstract: Neural networks are appealing for many applications since they are able to learn complex non-linear relationships between input and output data. As the number of neurons and layers in a neural network increase, it is possible to represent more complex relationships with automatically extracted features. Nowadays Deep Neural Networks (DNNs) are widely used in Computer Vision problems such as; classification, object detection, segmentation image editing etc. In this work, Facial Emotion Recognition task is performed by proposed Convolutional Neural Network (CNN)-based DNN architecture using FER2013 Dataset. Moreover, the effects of different hyperparameters (activation function, kernel size, initializer, batch size and network size) are investigated and ablation study results for Pooling Layer, Dropout and Batch Normalization are presented.
Abstract: The use of deep learning for species identification in camera trap images has revolutionised our ability to study, conserve and monitor species in a highly efficient and unobtrusive manner, with state-of-the-art models achieving accuracies surpassing the accuracy of manual human classification. The high imbalance of camera trap datasets, however, results in poor accuracies for minority (rare or endangered) species due to their relative insignificance to the overall model accuracy. This paper investigates the use of Focal Loss, in comparison to the traditional Cross Entropy Loss function, to improve the identification of minority species in the “255 Bird Species” dataset from Kaggle. The results show that, although Focal Loss slightly decreased the accuracy of the majority species, it was able to increase the F1-score by 0.06 and improve the identification of the bottom two, five and ten (minority) species by 37.5%, 15.7% and 10.8%, respectively, as well as resulting in an improved overall accuracy of 2.96%.
Abstract: With the rapid development of deep learning, neural network and deep learning algorithms play a significant role in various practical applications. Due to the high accuracy and good performance, Convolutional Neural Networks (CNNs) especially have become a research hot spot in the past few years. However, the size of the networks becomes increasingly large scale due to the demands of the practical applications, which poses a significant challenge to construct a high-performance implementation of deep learning neural networks. Meanwhile, many of these application scenarios also have strict requirements on the performance and low-power consumption of hardware devices. Therefore, it is particularly critical to choose a moderate computing platform for hardware acceleration of CNNs. This article aimed to survey the recent advance in Field Programmable Gate Array (FPGA)-based acceleration of CNNs. Various designs and implementations of the accelerator based on FPGA under different devices and network models are overviewed, and the versions of Graphic Processing Units (GPUs), Application Specific Integrated Circuits (ASICs) and Digital Signal Processors (DSPs) are compared to present our own critical analysis and comments. Finally, we give a discussion on different perspectives of these acceleration and optimization methods on FPGA platforms to further explore the opportunities and challenges for future research. More helpfully, we give a prospect for future development of the FPGA-based accelerator.
Abstract: The reason for conducting this research is to develop an algorithm that is capable of classifying news articles from the automobile industry, according to the competitive actions that they entail, with the use of Text Mining (TM) methods. It is needed to test how to properly preprocess the data for this research by preparing pipelines which fits each algorithm the best. The pipelines are tested along with nine different classification algorithms in the realm of regression, support vector machines, and neural networks. Preliminary testing for identifying the optimal pipelines and algorithms resulted in the selection of two algorithms with two different pipelines. The two algorithms are Logistic Regression (LR) and Artificial Neural Network (ANN). These algorithms are optimized further, where several parameters of each algorithm are tested. The best result is achieved with the ANN. The final model yields an accuracy of 0.79, a precision of 0.80, a recall of 0.78, and an F1 score of 0.76. By removing three of the classes that created noise, the final algorithm is capable of reaching an accuracy of 94%.
Abstract: Telecommunication service providers demand accurate
and precise prediction of customer churn probabilities to increase the
effectiveness of their customer relation services. The large amount of
customer data owned by the service providers is suitable for analysis
by machine learning methods. In this study, expenditure data of
customers are analyzed by using an artificial neural network (ANN).
The ANN model is applied to the data of customers with different
billing duration. The proposed model successfully predicts the churn
probabilities at 83% accuracy for only three months expenditure data
and the prediction accuracy increases up to 89% when the nine month
data is used. The experiments also show that the accuracy of ANN
model increases on an extended feature set with information of the
changes on the bill amounts.