Search

Accurate Origin-Destination (OD) prediction is significant for effective traffic monitor, which can support operation decision in traffic planning and management field. The enclosed expressway network system like toll gates system in China can collect mounts of trip records which can be gathered for OD prediction. The paper develops a novel neural network, which is named Expressway OD Prediction Neural Network (EODPNN) for toll data-based prediction. The network consists of the following three modules: The Feature Extension Module, the Memory Module, and the Prediction Module. In the process, the attributes data which can reflect the city attribute such as GDP, population, and the number of vehicles are considered to embeded into the notwork to increase the accuracy of the model. For the applicability improvment of the model, we categorize the cities in multiple classes based on their economy and population scales in this paper, which can provide a higher accurate prediction of OD by EODPNN. The results shows that, comparing to the traditional model like ARIMA and SVM, or typical neural networks like Bidirectional Long Short-term Memory, the EODPNN delivers a better prediction performance. The method proposed in this paper has been fully verified and has a potential to transplant to the other OD data-based management systems for a more accurate and flexible prediction. [ABSTRACT FROM AUTHOR]

Past research on the process of extinguishing a fire typically used a traditional linear water jet falling point model and the results ignored external factors, such as environmental conditions and the status of the fire engine, even though the water jet falling point location prediction was often associated with these parameters and showed a nonlinear relationship. This paper constructed a BP (Back Propagation) neural network model. The fire gun nozzle characteristics were included as model inputs, and the water discharge point coordinates were the model outputs; thus, the model could precisely predict the water discharge point with small error and high precision to determine an accurate firing position and allow for the timely adjustment of the spray gun. To improve the slow convergence and local optimality problems of the BP neural network (BPNN), this paper further used a genetic algorithm to optimize the BPNN (GA-BPNN). The BPNN can be used to optimize the weights in the network to train them for global optimization. A genetic algorithm was introduced into the neural network approach, and the water jet landing prediction model was further improved. The simulation results showed that the prediction accuracy of the GA-BP model was better than that of the BPNN alone. The established model can accurately predict the location of the water jet, making the prediction results more useful for firefighters. [ABSTRACT FROM AUTHOR]

Classifying bank accounts by using transaction data is encouraging in cracking down on illegal financial activities. However, few research simultaneously use heterogenous features, which are embedded in the time series data. In this paper, a two route convolution neural network TRHD-CNN model, fed with two types of heterogeneous feature matrices, is proposed for classifying the bank accounts. TRHD-CNN adopts divide and conquer strategy to extract characteristics from two types of data source independently. The strategy is proved able in mining complementary classification characteristics. We firstly transfer the original log data into a directed and dynamic transaction network. On the basis of that, two feature generation methods are devised for extracting information from local topological structure and time series transaction respectively. A DirectedWalk method is developed in this paper to learning the network vector of vertices used for embedding the neighbor relationship of bank account. The extensive experimental results, conducted on a real bank transaction dataset that contains illegal pyramid selling accounts, show the significant advantage of TRHD-CNN over the existing methods. TRHD-CNN can provide recall scores up to 5.15% higher than competing methods. In addition, the two-route architecture of TRHD-CNN is easy to extend to multi-route scenarios and other fields. [ABSTRACT FROM AUTHOR]

Analyzing and predicting the trend of price fluctuation has been receiving more and more attention, as price risk has become the focus of risk control research in heating oil futures market. A novel time series prediction model combined with the complex network method is put forward in the paper. First of all, this paper counts the cumulative time interval of different nodes in the network, and fits its growth trend with the Fourier model. Then a novel price fluctuation prediction model is established based on the effective information such as some topology properties extracted from the network. The results show that the Fourier model can predict the emergence time of new nodes in the next stage, and the established price fluctuation prediction model can infer the names of nodes in the prediction interval, so as to determine the forward-looking behavior of price evolution. Besides, liken to the NAR neural network, the prediction results obtained by the proposed method also show superiority, which has important theoretical value and academic significance for early warning and prediction of price behavior in the heating oil futures market. [ABSTRACT FROM AUTHOR]

Combining a deep neural network with fuzzy theory, this paper proposes an advertising click-through rate (CTR) prediction approach based on a fuzzy deep neural network (FDNN). In this approach, fuzzy Gaussian-Bernoulli restricted Boltzmann machine (FGBRBM) is first applied to input raw data from advertising datasets. Next, fuzzy restricted Boltzmann machine (FRBM) is used to construct the fuzzy deep belief network (FDBN) with the unsupervised method layer by layer. Finally, fuzzy logistic regression (FLR) is utilized for modeling the CTR. The experimental results show that the proposed FDNN model outperforms several baseline models in terms of both data representation capability and robustness in advertising click log datasets with noise. [ABSTRACT FROM AUTHOR]

In this paper, we propose a hybrid method to forecast the stock prices called High-order-fuzzy-fluctuation-Trends-based Back Propagation(HTBP)Neural Network model. First, we compare each value of the historical training data with the previous day's value to obtain a fluctuation trend time series (FTTS). On this basis, the FTTS blur into fuzzy time series (FFTS) based on the fluctuation of the increasing, equality, decreasing amplitude and direction. Since the relationship between FFTS and future wave trends is nonlinear, the HTBP neural network algorithm is used to find the mapping rules in the form of self-learning. Finally, the results of the algorithm output are used to predict future fluctuations. The proposed model provides some innovative features:(1)It combines fuzzy set theory and neural network algorithm to avoid overfitting problems existed in traditional models. (2)BP neural network algorithm can intelligently explore the internal rules of the actual existence of sequential data, without the need to analyze the influence factors of specific rules and the path of action. (3)The hybrid modal can reasonably remove noises from the internal rules by proper fuzzy treatment. This paper takes the TAIEX data set of Taiwan stock exchange as an example, and compares and analyzes the prediction performance of the model. The experimental results show that this method can predict the stock market in a very simple way. At the same time, we use this method to predict the Shanghai stock exchange composite index, and further verify the effectiveness and universality of the method. [ABSTRACT FROM AUTHOR]

This paper mainly studies the globally fixed-time synchronization of a class of coupled neutral-type neural networks with mixed time-varying delays via discontinuous feedback controllers. Compared with the traditional neutral-type neural network model, the model in this paper is more general. A class of general discontinuous feedback controllers are designed. With the help of the definition of fixed-time synchronization, the upper right-hand derivative and a defined simple Lyapunov function, some easily verifiable and extensible synchronization criteria are derived to guarantee the fixed-time synchronization between the drive and response systems. Finally, two numerical simulations are given to verify the correctness of the results. [ABSTRACT FROM AUTHOR]

Background: In this paper we present a model of parameters to aesthetically characterize films using a multi-disciplinary approach: by combining film theory, visual low-level video descriptors (modeled in order to supply aesthetic information) and classification techniques using machine and deep learning. Methods: Four different tests have been developed, each for a different application, proving the model's usefulness. These applications are: aesthetic style clustering, prediction of production year, genre detection and influence on film popularity. Results: The results are compared against high-level information to determine the accuracy of the model to classify films without knowing such information previously. The main difference with other film characterization approaches is that we are able to isolate the influence of high-level descriptors to really understand the relevance of low-level features and, accordingly propose a useful set of low-level visual descriptors for that purpose. This model has been tested with a representative number of films to prove that it can be used for different applications. [ABSTRACT FROM AUTHOR]

In this paper we investigate the changes in the functional connectivity intensity, and some related properties, in healthy people, across the life span and at resting state. For the explicit computation of the functional connectivity we exploit a recently proposed model, that bases not only on the correlations data provided by the acquisition equipment, but also on different parameters, such as the anatomical distances between nodes and their degrees. The leading purpose of the paper is to show that the proposed approach is able to recover the main aspects of resting state condition known from the available literature, as well as to suggest new insights, perspectives and speculations from a neurobiological point of view. Our study involves 133 subjects, both males and females of different ages, with no evidence of neurological diseases or systemic disorders. First, we show how the model applies to the sample, where the subjects are grouped into 28 different groups (14 of males and 14 of females), according to their age. This leads to the construction of two graphs (one for males and one for females), that can be realistically interpreted as representative of the neural network during the resting state. Second, following the idea that the brain network is better understood by focusing on specific nodes having a kind of centrality, we refine the two output graphs by introducing a new metric that favours the selection of nodes having higher degrees. As a third step, we extensively comment and discuss the obtained results. In particular, it is remarkable that, despite a great overlapping exists between the outcomes concerning males and females, some intriguing differences appear. This motivates a deeper local investigation, which represents the fourth part of the paper, carried out through a thorough statistical analysis. As a result, we are enabled to support that, for two special age groups, a few links contribute in differentiating the behaviour of males and females. In addition, we performed an average-based comparison between the proposed model and the traditional statistical correlation-based approach, then discussing and commenting the main outlined discrepancies. [ABSTRACT FROM AUTHOR]

This paper develops a new machine vision framework for efficient detection and classification of manufacturing defects in metal boxes. Previous techniques, which are based on either visual inspection or on hand-crafted features, are both inaccurate and time consuming. In this paper, we show that by using autoencoder deep neural network (DNN) architecture, we are able to not only classify manufacturing defects, but also localize them with high accuracy. Compared to traditional techniques, DNNs are able to learn, in a supervised manner, the visual features that achieve the best performance. Our experiments on a database of real images demonstrate that our approach overcomes the state-of-the-art while remaining computationally competitive. [ABSTRACT FROM AUTHOR]

Classical definitions of observability classify a system as either being observable or not. Observability has been recognized as an important feature to study complex networks, and as for dynamical systems the focus has been on determining conditions for a network to be observable. About twenty years ago continuous measures of observability for nonlinear dynamical systems started to be used. In this paper various aspects of observability that are established for dynamical systems will be investigated in the context of networks. In particular it will be discussed in which ways simple networks can be ranked in terms of observability using continuous measures of such a property. Also it is pointed out that the analysis of the network topology is typically not sufficient for observability purposes, since both the dynamics and the coupling of such nodes play a vital role. Some of the main ideas are illustrated by means of numerical simulations. [ABSTRACT FROM AUTHOR]

In recent years, finding the cause of pathogenesis is expected by observing the cell images. In this paper, we propose a cell particle detection method in cell images. However, there are mainly two kinds of problems in particle detection in cell image. The first is the different properties between cell images and standard images used in computer vision researches. Edges of cell particles are ambiguous, and overlaps between cell particles are often occurred in dense regions. It is difficult to detect cell particles by simple detection method using a binary classifier. The second is the ground truth made by cell biologists. The number of training samples for training a classifier is limited, and incorrect samples are included by the subjectivity of observers. From the background, we propose a cell particle detection method to address those problems. In our proposed method, we predict the center of a cell particle from the peripheral regions by convolutional neural network, and the prediction results are voted. By using the obvious peripheral edges, we can robustly detect overlapped cell particles because all edges of overlapping cell particles are not ambiguous. In addition, voting from peripheral views enables reliable detection. Moreover, our method is useful in practical applications because we can prepare many training samples from a cell particle. In experiments, we evaluate our detection methods on two kinds of cell detection datasets. One is challenging dataset for synthetic cells, and our method achieved the state-of-the-art performance. The other is real dataset of lipid droplets, and our method outperformed the conventional detector using CNN with binary outputs for particles and non-particles classification. [ABSTRACT FROM AUTHOR]

In palynology, the visual classification of pollen grains from different species is a hard task which is usually tackled by human operators using microscopes. Its complete automatization would save a high quantity of resources and provide valuable improvements especially for allergy-related information systems, but also for other application fields as paleoclimate reconstruction, quality control of honey based products, collection of evidences in criminal investigations or fabric dating and tracking. This paper presents three state-of-the-art deep learning classification methods applied to the recently published POLEN23E image dataset. The three methods make use of convolutional neural networks: the first one is strictly based on the idea of transfer learning, the second one is based on feature extraction and the third one represents a hybrid approach, combining transfer learning and feature extraction. The results from the three methods are indeed very good, reaching over 97% correct classification rates in images not previously seen by the models, where other authors reported around 70. [ABSTRACT FROM AUTHOR]

Textual representations play an important role in the field of natural language processing (NLP). The efficiency of NLP tasks, such as text comprehension and information extraction, can be significantly improved with proper textual representations. As neural networks are gradually applied to learn the representation of words and phrases, fairly efficient models of learning short text representations have been developed, such as the continuous bag of words (CBOW) and skip-gram models, and they have been extensively employed in a variety of NLP tasks. Because of the complex structure generated by the longer text lengths, such as sentences, algorithms appropriate for learning short textual representations are not applicable for learning long textual representations. One method of learning long textual representations is the Long Short-Term Memory (LSTM) network, which is suitable for processing sequences. However, the standard LSTM does not adequately address the primary sentence structure (subject, predicate and object), which is an important factor for producing appropriate sentence representations. To resolve this issue, this paper proposes the dependency-based LSTM model (D-LSTM). The D-LSTM divides a sentence representation into two parts: a basic component and a supporting component. The D-LSTM uses a pre-trained dependency parser to obtain the primary sentence information and generate supporting components, and it also uses a standard LSTM model to generate the basic sentence components. A weight factor that can adjust the ratio of the basic and supporting components in a sentence is introduced to generate the sentence representation. Compared with the representation learned by the standard LSTM, the sentence representation learned by the D-LSTM contains a greater amount of useful information. The experimental results show that the D-LSTM is superior to the standard LSTM for sentences involving compositional knowledge (SICK) data. [ABSTRACT FROM AUTHOR]

Authorship attribution is to identify the most likely author of a given sample among a set of candidate known authors. It can be not only applied to discover the original author of plain text, such as novels, blogs, emails, posts etc., but also used to identify source code programmers. Authorship attribution of source code is required in diverse applications, ranging from malicious code tracking to solving authorship dispute or software plagiarism detection. This paper aims to propose a new method to identify the programmer of Java source code samples with a higher accuracy. To this end, it first introduces back propagation (BP) neural network based on particle swarm optimization (PSO) into authorship attribution of source code. It begins by computing a set of defined feature metrics, including lexical and layout metrics, structure and syntax metrics, totally 19 dimensions. Then these metrics are input to neural network for supervised learning, the weights of which are output by PSO and BP hybrid algorithm. The effectiveness of the proposed method is evaluated on a collected dataset with 3,022 Java files belong to 40 authors. Experiment results show that the proposed method achieves 91.060% accuracy. And a comparison with previous work on authorship attribution of source code for Java language illustrates that this proposed method outperforms others overall, also with an acceptable overhead. [ABSTRACT FROM AUTHOR]

Modern computing architecture based on the separation of memory and processing leads to a well known problem called the von Neumann bottleneck, a restrictive limit on the data bandwidth between CPU and RAM. This paper introduces a new approach to computing we call AHaH computing where memory and processing are combined. The idea is based on the attractor dynamics of volatile dissipative electronics inspired by biological systems, presenting an attractive alternative architecture that is able to adapt, self-repair, and learn from interactions with the environment. We envision that both von Neumann and AHaH computing architectures will operate together on the same machine, but that the AHaH computing processor may reduce the power consumption and processing time for certain adaptive learning tasks by orders of magnitude. The paper begins by drawing a connection between the properties of volatility, thermodynamics, and Anti-Hebbian and Hebbian (AHaH) plasticity. We show how AHaH synaptic plasticity leads to attractor states that extract the independent components of applied data streams and how they form a computationally complete set of logic functions. After introducing a general memristive device model based on collections of metastable switches, we show how adaptive synaptic weights can be formed from differential pairs of incremental memristors. We also disclose how arrays of synaptic weights can be used to build a neural node circuit operating AHaH plasticity. By configuring the attractor states of the AHaH node in different ways, high level machine learning functions are demonstrated. This includes unsupervised clustering, supervised and unsupervised classification, complex signal prediction, unsupervised robotic actuation and combinatorial optimization of procedures–all key capabilities of biological nervous systems and modern machine learning algorithms with real world application. [ABSTRACT FROM AUTHOR]

We propose the architecture of a novel robot system merging biological and artificial intelligence based on a neural controller connected to an external agent. We initially built a framework that connected the dissociated neural network to a mobile robot system to implement a realistic vehicle. The mobile robot system characterized by a camera and two-wheeled robot was designed to execute the target-searching task. We modified a software architecture and developed a home-made stimulation generator to build a bi-directional connection between the biological and the artificial components via simple binomial coding/decoding schemes. In this paper, we utilized a specific hierarchical dissociated neural network for the first time as the neural controller. Based on our work, neural cultures were successfully employed to control an artificial agent resulting in high performance. Surprisingly, under the tetanus stimulus training, the robot performed better and better with the increasement of training cycle because of the short-term plasticity of neural network (a kind of reinforced learning). Comparing to the work previously reported, we adopted an effective experimental proposal (i.e. increasing the training cycle) to make sure of the occurrence of the short-term plasticity, and preliminarily demonstrated that the improvement of the robot’s performance could be caused independently by the plasticity development of dissociated neural network. This new framework may provide some possible solutions for the learning abilities of intelligent robots by the engineering application of the plasticity processing of neural networks, also for the development of theoretical inspiration for the next generation neuro-prostheses on the basis of the bi-directional exchange of information within the hierarchical neural networks. [ABSTRACT FROM AUTHOR]

Stock price prediction is an important and challenging problem in stock market analysis. Existing prediction methods either exploit autocorrelation of stock price and its correlation with the supply and demand of stock, or explore predictive indictors exogenous to stock market. In this paper, using transaction record of stocks with identifier of traders, we introduce an index to characterize market confidence, i.e., the ratio of the number of traders who is active in two successive trading days to the number of active traders in a certain trading day. Strong Granger causality is found between the index of market confidence and stock price. We further predict stock price by incorporating the index of market confidence into a neural network based on time series of stock price. Experimental results on 50 stocks in two Chinese Stock Exchanges demonstrate that the accuracy of stock price prediction is significantly improved by the inclusion of the market confidence index. This study sheds light on using cross-day trading behavior to characterize market confidence and to predict stock price. [ABSTRACT FROM AUTHOR]

Intelligent Transportation Systems rely on understanding, predicting and affecting the interactions between vehicles. The goal of this paper is to choose a small subset from the larger set so that the resulting regression model is simple, yet have good predictive ability for Vehicle agent speed relative to Vehicle intruder. The method of ANFIS (adaptive neuro fuzzy inference system) was applied to the data resulting from these measurements. The ANFIS process for variable selection was implemented in order to detect the predominant variables affecting the prediction of agent speed relative to intruder. This process includes several ways to discover a subset of the total set of recorded parameters, showing good predictive capability. The ANFIS network was used to perform a variable search. Then, it was used to determine how 9 parameters (Intruder Front sensors active (boolean), Intruder Rear sensors active (boolean), Agent Front sensors active (boolean), Agent Rear sensors active (boolean), RSSI signal intensity/strength (integer), Elapsed time (in seconds), Distance between Agent and Intruder (m), Angle of Agent relative to Intruder (angle between vehicles °), Altitude difference between Agent and Intruder (m)) influence prediction of agent speed relative to intruder. The results indicated that distance between Vehicle agent and Vehicle intruder (m) and angle of Vehicle agent relative to Vehicle Intruder (angle between vehicles °) is the most influential parameters to Vehicle agent speed relative to Vehicle intruder. [ABSTRACT FROM AUTHOR]

Swarming behavior is common in biology, from cell colonies to insect swarms and bird flocks. However, the conditions leading to the emergence of such behavior are still subject to research. Since Reynolds’ boids, many artificial models have reproduced swarming behavior, focusing on details ranging from obstacle avoidance to the introduction of fixed leaders. This paper presents a model of evolved artificial agents, able to develop swarming using only their ability to listen to each other’s signals. The model simulates a population of agents looking for a vital resource they cannot directly detect, in a 3D environment. Instead of a centralized algorithm, each agent is controlled by an artificial neural network, whose weights are encoded in a genotype and adapted by an original asynchronous genetic algorithm. The results demonstrate that agents progressively evolve the ability to use the information exchanged between each other via signaling to establish temporary leader-follower relations. These relations allow agents to form swarming patterns, emerging as a transient behavior that improves the agents’ ability to forage for the resource. Once they have acquired the ability to swarm, the individuals are able to outperform the non-swarmers at finding the resource. The population hence reaches a neutral evolutionary space which leads to a genetic drift of the genotypes. This reductionist approach to signal-based swarming not only contributes to shed light on the minimal conditions for the evolution of a swarming behavior, but also more generally it exemplifies the effect communication can have on optimal search patterns in collective groups of individuals. [ABSTRACT FROM AUTHOR]

The spiking neural networks (SNNs) are the third generation of neural networks and perform remarkably well in cognitive tasks such as pattern recognition. The spike emitting and information processing mechanisms found in biological cognitive systems motivate the application of the hierarchical structure and temporal encoding mechanism in spiking neural networks, which have exhibited strong computational capability. However, the hierarchical structure and temporal encoding approach require neurons to process information serially in space and time respectively, which reduce the training efficiency significantly. For training the hierarchical SNNs, most existing methods are based on the traditional back-propagation algorithm, inheriting its drawbacks of the gradient diffusion and the sensitivity on parameters. To keep the powerful computation capability of the hierarchical structure and temporal encoding mechanism, but to overcome the low efficiency of the existing algorithms, a new training algorithm, the Normalized Spiking Error Back Propagation (NSEBP) is proposed in this paper. In the feedforward calculation, the output spike times are calculated by solving the quadratic function in the spike response model instead of detecting postsynaptic voltage states at all time points in traditional algorithms. Besides, in the feedback weight modification, the computational error is propagated to previous layers by the presynaptic spike jitter instead of the gradient decent rule, which realizes the layer-wised training. Furthermore, our algorithm investigates the mathematical relation between the weight variation and voltage error change, which makes the normalization in the weight modification applicable. Adopting these strategies, our algorithm outperforms the traditional SNN multi-layer algorithms in terms of learning efficiency and parameter sensitivity, that are also demonstrated by the comprehensive experimental results in this paper. [ABSTRACT FROM AUTHOR]

Short-term metro passenger flow forecasting is an essential component of intelligent transportation systems (ITS) and can be applied to optimize the passenger flow organization of a station and offer data support for metro passenger flow early warning and system management. LSTM neural networks have recently achieved remarkable recent in the field of natural language processing (NLP) because they are well suited for learning from experience to predict time series. For this purpose, we propose an empirical mode decomposition (EMD)-based long short-term memory (LSTM) neural network model for predicting short-term metro inbound passenger flow. The EMD algorithm decomposes the original sequential passenger flow into several intrinsic mode functions (IMFs) and a residual. Selected IMFs that are strongly correlated with the original data can be obtained via feature selection. The selected IMFs and the original data are integrated into inputs for LSTM neural networks, and a single LSTM prediction model and an EMD-LSTM hybrid forecasting model are developed. Finally, historical real automatic fare collection (AFC) data from metro passengers are collected from Chengdu Metro to verify the validity of the proposed EMD-LSTM prediction model. The results indicate that the proposed EMD-LSTM hybrid forecasting model outperforms the LSTM, ARIMA and BPN models. [ABSTRACT FROM AUTHOR]

Single-point incremental forming (SPIF) is a technology that allows incremental manufacturing of complex parts from a flat sheet using simple tools; further, this technology is flexible and economical. Measuring the forming force using this technology helps in preventing failures, determining the optimal processes, and implementing on-line control. In this paper, an experimental study using SPIF is described. This study focuses on the influence of four different process parameters, namely, step size, tool diameter, sheet thickness, and feed rate, on the maximum forming force. For an efficient force predictive model based on an adaptive neuro-fuzzy inference system (ANFIS), an artificial neural network (ANN) and a regressions model were applied. The predicted forces exhibited relatively good agreement with the experimental results. The results indicate that the performance of the ANFIS model realizes the full potential of the ANN model. [ABSTRACT FROM AUTHOR]

Locating diseases precisely from medical images, like ultrasonic and CT images, have been one of the most challenging problems in medical image analysis. In recent years, the vigorous development of deep learning models have greatly improved the accuracy in disease location on medical images. However, there are few artificial intelligent methods for identifying cholelithiasis and classifying gallstones on CT images, since no open source CT images dataset of cholelithiasis and gallstones is available for training the models and verifying their performance. In this paper, we build up the first medical image dataset of cholelithiasis by collecting 223846 CT images with gallstone of 1369 patients. With these CT images, a neural network is trained to “pick up” CT images of high quality as training set, and then a novel Yolo neural network, named Yolov3-arch neural network, is proposed to identify cholelithiasis and classify gallstones on CT images. Identification and classification accuracies are obtained by 10-fold cross-validations. It is obtained that our Yolov3-arch model is with average accuracy 92.7% in identifying granular gallstones and average accuracy 80.3% in identifying muddy gallstones. This achieves 3.5% and 8% improvements in identifying granular and muddy gallstones to general Yolo v3 model, respectively. Also, the average cholelithiasis identifying accuracy is improved to 86.50% from 80.75%. Meanwhile, our method can reduce the misdiagnosis rate of negative samples by the object detection model. [ABSTRACT FROM AUTHOR]

Skin cancer is one of most deadly diseases in humans. According to the high similarity between melanoma and nevus lesions, physicians take much more time to investigate these lesions. The automated classification of skin lesions will save effort, time and human life. The purpose of this paper is to present an automatic skin lesions classification system with higher classification rate using the theory of transfer learning and the pre-trained deep neural network. The transfer learning has been applied to the Alex-net in different ways, including fine-tuning the weights of the architecture, replacing the classification layer with a softmax layer that works with two or three kinds of skin lesions, and augmenting dataset by fixed and random rotation angles. The new softmax layer has the ability to classify the segmented color image lesions into melanoma and nevus or into melanoma, seborrheic keratosis, and nevus. The three well-known datasets, MED-NODE, Derm (IS & Quest) and ISIC, are used in testing and verifying the proposed method. The proposed DCNN weights have been fine-tuned using the training and testing dataset from ISIC in addition to 10-fold cross validation for MED-NODE and DermIS—DermQuest. The accuracy, sensitivity, specificity, and precision measures are used to evaluate the performance of the proposed method and the existing methods. For the datasets, MED-NODE, Derm (IS & Quest) and ISIC, the proposed method has achieved accuracy percentages of 96.86%, 97.70%, and 95.91% respectively. The performance of the proposed method has outperformed the performance of the existing classification methods of skin cancer. [ABSTRACT FROM AUTHOR]

Electroencephalogram (EEG) is a common base signal used to monitor brain activities and diagnose sleep disorders. Manual sleep stage scoring is a time-consuming task for sleep experts and is limited by inter-rater reliability. In this paper, we propose an automatic sleep stage annotation method called SleepEEGNet using a single-channel EEG signal. The SleepEEGNet is composed of deep convolutional neural networks (CNNs) to extract time-invariant features, frequency information, and a sequence to sequence model to capture the complex and long short-term context dependencies between sleep epochs and scores. In addition, to reduce the effect of the class imbalance problem presented in the available sleep datasets, we applied novel loss functions to have an equal misclassified error for each sleep stage while training the network. We evaluated the performance of the proposed method on different single-EEG channels (i.e., Fpz-Cz and Pz-Oz EEG channels) from the Physionet Sleep-EDF datasets published in 2013 and 2018. The evaluation results demonstrate that the proposed method achieved the best annotation performance compared to current literature, with an overall accuracy of 84.26%, a macro F1-score of 79.66% and κ = 0.79. Our developed model can be applied to other sleep EEG signals and aid the sleep specialists to arrive at an accurate diagnosis. The source code is available at . [ABSTRACT FROM AUTHOR]

Antenna selection in Multiple-Input Multiple-Output (MIMO) systems has attracted increasing attention due to the challenge of keeping a balance between communication performance and computational complexity. Recently, deep learning based methods have achieved promising performance in many application fields. This paper proposed a deep learning (DL) based antenna selection technique. First, we generated the label of training antenna systems by maximizing the channel capacity. Then, we adopted the deep convolutional neural network (CNN) on the channel matrices to explicitly exploit the massive latent cues of attenuation coefficients. Finally, we used the adopted CNN to assign the class label and then select the optimal antenna subset. Experimental results demonstrate that our method can achieve better performance than the state-of-the-art baselines for data-driven based antenna selection. [ABSTRACT FROM AUTHOR]

Although successful detection of malignant tumors from histopathological images largely depends on the long-term experience of radiologists, experts sometimes disagree with their decisions. Computer-aided diagnosis provides a second option for image diagnosis, which can improve the reliability of experts’ decision-making. Automatic and precision classification for breast cancer histopathological image is of great importance in clinical application for identifying malignant tumors from histopathological images. Advanced convolution neural network technology has achieved great success in natural image classification, and it has been used widely in biomedical image processing. In this paper, we design a novel convolutional neural network, which includes a convolutional layer, small SE-ResNet module, and fully connected layer. We propose a small SE-ResNet module which is an improvement on the combination of residual module and Squeeze-and-Excitation block, and achieves the similar performance with fewer parameters. In addition, we propose a new learning rate scheduler which can get excellent performance without complicatedly fine-tuning the learning rate. We use our model for the automatic classification of breast cancer histology images (BreakHis dataset) into benign and malignant and eight subtypes. The results show that our model achieves the accuracy between 98.87% and 99.34% for the binary classification and achieve the accuracy between 90.66% and 93.81% for the multi-class classification. [ABSTRACT FROM AUTHOR]

Computer-aided polyp detection in gastric gastroscopy has been the subject of research over the past few decades. However, despite significant advances, automatic polyp detection in real time is still an unsolved problem. In this paper, we report on a convolutional neural network (CNN) for polyp detection that is constructed based on Single Shot MultiBox Detector (SSD) architecture and which we call SSD for Gastric Polyps (SSD-GPNet). To take full advantages of feature maps’ information from the feature pyramid and to acquire higher accuracy, we re-use information that is abandoned by Max-Pooling layers. In other words, we reuse the lost data from the pooling layers and concatenate that data as extra feature maps to contribute to classification and detection. Meanwhile, in the feature pyramid, we concatenate feature maps of the lower layers and feature maps that are deconvolved from upper layers to make explicit relationships between layers and to effectively increase the number of channels. The results show that our enhanced SSD for gastric polyp detection can realize real-time polyp detection with 50 frames per second (FPS) and can improve the mean average precision (mAP) from 88.5% to 90.4%, with only a little loss in time-performance. And the further experiment shows that SSD-GPNet has excellent performance in improving polyp detection recalls over 10% (p = 0.00053), especially in small polyp detection. This can help endoscopic physicians more easily find missed polyps and decrease the gastric polyp miss rate. It may be applicable in daily clinical practice to reduce the burden on physicians. [ABSTRACT FROM AUTHOR]

Recognizing a person in a crowded environment is a challenging, yet critical, visual-search task for both humans and machine-vision algorithms. This paper explores the possibility of combining a residual neural network (ResNet), brain-computer interfaces (BCIs) and human participants to create “cyborgs” that improve decision making. Human participants and a ResNet undertook the same face-recognition experiment. BCIs were used to decode the decision confidence of humans from their EEG signals. Different types of cyborg groups were created, including either only humans (with or without the BCI) or groups of humans and the ResNet. Cyborg groups decisions were obtained weighing individual decisions by confidence estimates. Results show that groups of cyborgs are significantly more accurate (up to 35%) than the ResNet, the average participant, and equally-sized groups of humans not assisted by technology. These results suggest that melding humans, BCI, and machine-vision technology could significantly improve decision-making in realistic scenarios. [ABSTRACT FROM AUTHOR]

It is evident through biology research that, biological neural network could be implemented through two means: by congenital heredity, or by posteriority learning. However, traditionally, artificial neural network, especially the Deep learning Neural Networks (DNNs) are implemented only through exhaustive training and learning. Fixed structure is built, and then parameters are trained through huge amount of data. In this way, there are a lot of redundancies in the implemented artificial neural network. This redundancy not only requires more effort to train the network, but also costs more computing resources when used. In this paper, we proposed a bionic way to implement artificial neural network through construction rather than training and learning. The hierarchy of the neural network is designed according to analysis of the required functionality, and then module design is carried out to form each hierarchy. We choose the Drosophila’s visual neural network as a test case to verify our method’s validation. The results show that the bionic artificial neural network built through our method could work as a bionic compound eye, which can achieve the detection of the object and their movement, and the results are better on some properties, compared with the Drosophila’s biological compound eyes. [ABSTRACT FROM AUTHOR]

Health care organizations are leveraging machine-learning techniques, such as artificial neural networks (ANN), to improve delivery of care at a reduced cost. Applications of ANN to diagnosis are well-known; however, ANN are increasingly used to inform health care management decisions. We provide a seminal review of the applications of ANN to health care organizational decision-making. We screened 3,397 articles from six databases with coverage of Health Administration, Computer Science and Business Administration. We extracted study characteristics, aim, methodology and context (including level of analysis) from 80 articles meeting inclusion criteria. Articles were published from 1997–2018 and originated from 24 countries, with a plurality of papers (26 articles) published by authors from the United States. Types of ANN used included ANN (36 articles), feed-forward networks (25 articles), or hybrid models (23 articles); reported accuracy varied from 50% to 100%. The majority of ANN informed decision-making at the micro level (61 articles), between patients and health care providers. Fewer ANN were deployed for intra-organizational (meso- level, 29 articles) and system, policy or inter-organizational (macro- level, 10 articles) decision-making. Our review identifies key characteristics and drivers for market uptake of ANN for health care organizational decision-making to guide further adoption of this technique. [ABSTRACT FROM AUTHOR]

In general, the stock prices of the same industry have a similar trend, but those of different industries do not. When investing in stocks of different industries, one should select the optimal model from lots of trading models for each industry because any model may not be suitable for capturing the stock trends of all industries. However, the study has not been carried out at present. In this paper, firstly we select 424 S&P 500 index component stocks (SPICS) and 185 CSI 300 index component stocks (CSICS) as the research objects from 2010 to 2017, divide them into 9 industries such as finance and energy respectively. Secondly, we apply 12 widely used machine learning algorithms to generate stock trading signals in different industries and execute the back-testing based on the trading signals. Thirdly, we use a non-parametric statistical test to evaluate whether there are significant differences among the trading performance evaluation indicators (PEI) of different models in the same industry. Finally, we propose a series of rules to select the optimal models for stock investment of every industry. The analytical results on SPICS and CSICS show that we can find the optimal trading models for each industry based on the statistical tests and the rules. Most importantly, the PEI of the best algorithms can be significantly better than that of the benchmark index and “Buy and Hold” strategy. Therefore, the algorithms can be used for making profits from industry stock trading. [ABSTRACT FROM AUTHOR]

Effort-aware just-in-time (JIT) defect prediction is to rank source code changes based on the likelihood of detects as well as the effort to inspect such changes. Accurate defect prediction algorithms help to find more defects with limited effort. To improve the accuracy of defect prediction, in this paper, we propose a deep learning based approach for effort-aware just-in-time defect prediction. The key idea of the proposed approach is that neural network and deep learning could be exploited to select useful features for defect prediction because they have been proved excellent at selecting useful features for classification and regression. First, we preprocess ten numerical metrics of code changes, and then feed them to a neural network whose output indicates how likely the code change under test contains bugs. Second, we compute the benefit cost ratio for each code change by dividing the likelihood by its size. Finally, we rank code changes according to their benefit cost ratio. Evaluation results on a well-known data set suggest that the proposed approach outperforms the state-of-the-art approaches on each of the subject projects. It improves the average recall and popt by 15.6% and 8.1%, respectively. [ABSTRACT FROM AUTHOR]

Attention maps have been fused in the VggNet structure (EAC-Net) [] and have shown significant improvement compared to that of the VggNet structure. However, in [], E-Net was designed based on the facial action unit (AU) center and for facial AU detection only. Thus, for the use of attention maps in every image type, this paper proposed a new convolutional neural network (CNN) structure, P_VggNet, comprising the following parts: P_Net and VggNet with 16 layers (VggNet-16). The generation approach of P_Net was designed, and the P_VggNet structure was proposed. To prove the efficiency of P_VggNet, we designed two experiments, which indicated that P_VggNet could more efficiently extract image features than VggNet-16. [ABSTRACT FROM AUTHOR]

Background: Detection of ossification areas of hand bones in X-ray images is an important task, e.g. as a preprocessing step in automated bone age estimation. Deep neural networks have emerged recently as de facto standard detection methods, but their drawback is the need of large annotated datasets. Finetuning pre-trained networks is a viable alternative, but it is not clear a priori if training with small annotated datasets will be successful, as it depends on the problem at hand. In this paper, we show that pre-trained networks can be utilized to produce an effective detector of ossification areas in pediatric X-ray images of hands. Methods and findings: A publicly available Faster R-CNN network, pre-trained on the COCO dataset, was utilized and finetuned with 240 manually annotated radiographs from the RSNA Pediatric Bone Age Challenge, which comprises over 14.000 pediatric radiographs. The validation is done on another 89 radiographs from the dataset and the performance is measured by Intersection-over-Union (IoU). To understand the effect of the data size on the pre-trained network, subsampling was applied to the training data and the training was repeated. Additionally, the network was trained from scratch without any pre-trained weights. Finally, to understand whether the trained model could be useful, we compared the inference of the network to an annotation of an expert radiologist. The finetuned network was able to achieve an average precision (mAP@0.5IoU) of 92.92 ± 1.93. Apart from the wrist region, all ossification areas were able to benefit from more data. In contrast, the network trained from scratch was not able to produce any correct results. When compared to the annotations of the expert radiologist, the network was able to localize the regions quite well, as the F1-Score was on average 91.85 ± 1.06. Conclusions: By finetuning a pre-trained deep neural network, with 240 annotated radiographs, we were able to successfully detect ossification areas in prediatric hand radiographs. [ABSTRACT FROM AUTHOR]

Forecasting naturally occurring phenomena is a common problem in many domains of science, and this has been addressed and investigated by many scientists. The importance of time series prediction stems from the fact that it has wide range of applications, including control systems, engineering processes, environmental systems and economics. From the knowledge of some aspects of the previous behaviour of the system, the aim of the prediction process is to determine or predict its future behaviour. In this paper, we consider a novel application of a higher order polynomial neural network architecture called Dynamic Ridge Polynomial Neural Network that combines the properties of higher order and recurrent neural networks for the prediction of physical time series. In this study, four types of signals have been used, which are; The Lorenz attractor, mean value of the AE index, sunspot number, and heat wave temperature. The simulation results showed good improvements in terms of the signal to noise ratio in comparison to a number of higher order and feedforward neural networks in comparison to the benchmarked techniques. [ABSTRACT FROM AUTHOR]

In this paper, the synchronization problem for a class of discrete-time complex-valued neural networks with time-varying delays is investigated. Compared with the previous work, the time delay and parameters are assumed to be time-varying. By separating the real part and imaginary part, the discrete-time model of complex-valued neural networks is derived. Moreover, by using the complex-valued Lyapunov-Krasovskii functional method and linear matrix inequality as tools, sufficient conditions of the synchronization stability are obtained. In numerical simulation, examples are presented to show the effectiveness of our method. [ABSTRACT FROM AUTHOR]

Deep learning techniques, e.g., Convolutional Neural Networks (CNNs), have been explosively applied to the research in the fields of information retrieval and natural language processing. However, few research efforts have addressed semantic indexing with deep learning. The use of semantic indexing in the biomedical literature has been limited for several reasons. For instance, MEDLINE citations contain a large number of semantic labels from automatically annotated MeSH terms, and for a great deal of the literature, only the information of the title and the abstract is readily available. In this paper, we propose a Boltzmann Convolutional neural network framework (B-CNN) for biomedicine semantic indexing. In our hybrid learning framework, the CNN can adaptively deal with features of documents that have sequence relationships, and can capture context information accordingly; the Deep Boltzmann Machine (DBM) merges global (the entity in each document) and local information through its training with undirected connections. Additionally, we have designed a hierarchical coarse to fine style indexing structure for learning and classifying documents, and a novel feature extension approach with word sequence embedding and Wikipedia categorization. Comparative experiments were conducted for semantic indexing of biomedical abstract documents; these experiments verified the encouraged performance of our B-CNN model. [ABSTRACT FROM AUTHOR]

This paper presents a new method to reduce the computational cost when using Neural Networks as Language Models, during recognition, in some particular scenarios. It is based on a Neural Network that considers input contexts of different length in order to ease the use of a fallback mechanism together with the precomputation of softmax normalization constants for these inputs. The proposed approach is empirically validated, showing their capability to emulate lower order N-grams with a single Neural Network. A machine translation task shows that the proposed model constitutes a good solution to the normalization cost of the output softmax layer of Neural Networks, for some practical cases, without a significant impact in performance while improving the system speed. [ABSTRACT FROM AUTHOR]

Precise detection of invasive cancer on whole-slide images (WSI) is a critical first step in digital pathology tasks of diagnosis and grading. Convolutional neural network (CNN) is the most popular representation learning method for computer vision tasks, which have been successfully applied in digital pathology, including tumor and mitosis detection. However, CNNs are typically only tenable with relatively small image sizes (200 × 200 pixels). Only recently, Fully convolutional networks (FCN) are able to deal with larger image sizes (500 × 500 pixels) for semantic segmentation. Hence, the direct application of CNNs to WSI is not computationally feasible because for a WSI, a CNN would require billions or trillions of parameters. To alleviate this issue, this paper presents a novel method, High-throughput Adaptive Sampling for whole-slide Histopathology Image analysis (HASHI), which involves: i) a new efficient adaptive sampling method based on probability gradient and quasi-Monte Carlo sampling, and, ii) a powerful representation learning classifier based on CNNs. We applied HASHI to automated detection of invasive breast cancer on WSI. HASHI was trained and validated using three different data cohorts involving near 500 cases and then independently tested on 195 studies from The Cancer Genome Atlas. The results show that (1) the adaptive sampling method is an effective strategy to deal with WSI without compromising prediction accuracy by obtaining comparative results of a dense sampling (∼6 million of samples in 24 hours) with far fewer samples (∼2,000 samples in 1 minute), and (2) on an independent test dataset, HASHI is effective and robust to data from multiple sites, scanners, and platforms, achieving an average Dice coefficient of 76%. [ABSTRACT FROM AUTHOR]

Recent research on speech segregation and music fingerprinting has led to improvements in speech segregation and music identification algorithms. Speech and music segregation generally involves the identification of music followed by speech segregation. However, music segregation becomes a challenging task in the presence of noise. This paper proposes a novel method of speech segregation for unlabelled stationary noisy audio signals using the deep belief network (DBN) model. The proposed method successfully segregates a music signal from noisy audio streams. A recurrent neural network (RNN)-based hidden layer segregation model is applied to remove stationary noise. Dictionary-based fisher algorithms are employed for speech classification. The proposed method is tested on three datasets (TIMIT, MIR-1K, and MusicBrainz), and the results indicate the robustness of proposed method for speech segregation. The qualitative and quantitative analysis carried out on three datasets demonstrate the efficiency of the proposed method compared to the state-of-the-art speech segregation and classification-based methods. [ABSTRACT FROM AUTHOR]

Advanced Metering Infrastructure (AMI) realizes a two-way communication of electricity data through by interconnecting with a computer network as the core component of the smart grid. Meanwhile, it brings many new security threats and the traditional intrusion detection method can’t satisfy the security requirements of AMI. In this paper, an intrusion detection system based on Online Sequence Extreme Learning Machine (OS-ELM) is established, which is used to detecting the attack in AMI and carrying out the comparative analysis with other algorithms. Simulation results show that, compared with other intrusion detection methods, intrusion detection method based on OS-ELM is more superior in detection speed and accuracy. [ABSTRACT FROM AUTHOR]

In secondary analysis of electronic health records, a crucial task consists in correctly identifying the patient cohort under investigation. In many cases, the most valuable and relevant information for an accurate classification of medical conditions exist only in clinical narratives. Therefore, it is necessary to use natural language processing (NLP) techniques to extract and evaluate these narratives. The most commonly used approach to this problem relies on extracting a number of clinician-defined medical concepts from text and using machine learning techniques to identify whether a particular patient has a certain condition. However, recent advances in deep learning and NLP enable models to learn a rich representation of (medical) language. Convolutional neural networks (CNN) for text classification can augment the existing techniques by leveraging the representation of language to learn which phrases in a text are relevant for a given medical condition. In this work, we compare concept extraction based methods with CNNs and other commonly used models in NLP in ten phenotyping tasks using 1,610 discharge summaries from the MIMIC-III database. We show that CNNs outperform concept extraction based methods in almost all of the tasks, with an improvement in F1-score of up to 26 and up to 7 percentage points in area under the ROC curve (AUC). We additionally assess the interpretability of both approaches by presenting and evaluating methods that calculate and extract the most salient phrases for a prediction. The results indicate that CNNs are a valid alternative to existing approaches in patient phenotyping and cohort identification, and should be further investigated. Moreover, the deep learning approach presented in this paper can be used to assist clinicians during chart review or support the extraction of billing codes from text by identifying and highlighting relevant phrases for various medical conditions. [ABSTRACT FROM AUTHOR]

Detecting drug-drug interactions (DDI) is important because information on DDIs can help prevent adverse effects from drug combinations. Since there are many new DDI-related papers published in the biomedical domain, manually extracting DDI information from the literature is a laborious task. However, text mining can be used to find DDIs in the biomedical literature. Among the recently developed neural networks, we use a Recursive Neural Network to improve the performance of DDI extraction. Our recursive neural network model uses a position feature, a subtree containment feature, and an ensemble method to improve the performance of DDI extraction. Compared with the state-of-the-art models, the DDI detection and type classifiers of our model performed 4.4% and 2.8% better, respectively, on the DDIExtraction Challenge’13 test data. We also validated our model on the PK DDI corpus that consists of two types of DDIs data: in vivo DDI and in vitro DDI. Compared with the existing model, our detection classifier performed 2.3% and 6.7% better on in vivo and in vitro data respectively. The results of our validation demonstrate that our model can automatically extract DDIs better than existing models. [ABSTRACT FROM AUTHOR]

Since absorption of X-ray radiation has the possibility of inducing cancerous, genetic and other diseases to patients, researches usually attempt to reduce the radiation dose. However, reduction of the radiation dose associated with CT scans will unavoidably increase the severity of noise and artifacts, which can seriously affect diagnostic confidence. Due to the outstanding performance of deep neural networks in image processing, in this paper, we proposed a Stacked Competitive Network (SCN) approach to noise reduction, which stacks several successive Competitive Blocks (CB). The carefully handcrafted design of the competitive blocks was inspired by the idea of multi-scale processing and improvement the network’s capacity. Qualitative and quantitative evaluations demonstrate the competitive performance of the proposed method in noise suppression, structural preservation, and lesion detection. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new automatic hyperparameter selection approach for determining the optimal network configuration (network structure and hyperparameters) for deep neural networks using particle swarm optimization (PSO) in combination with a steepest gradient descent algorithm. In the proposed approach, network configurations were coded as a set of real-number m-dimensional vectors as the individuals of the PSO algorithm in the search procedure. During the search procedure, the PSO algorithm is employed to search for optimal network configurations via the particles moving in a finite search space, and the steepest gradient descent algorithm is used to train the DNN classifier with a few training epochs (to find a local optimal solution) during the population evaluation of PSO. After the optimization scheme, the steepest gradient descent algorithm is performed with more epochs and the final solutions (pbest and gbest) of the PSO algorithm to train a final ensemble model and individual DNN classifiers, respectively. The local search ability of the steepest gradient descent algorithm and the global search capabilities of the PSO algorithm are exploited to determine an optimal solution that is close to the global optimum. We constructed several experiments on hand-written characters and biological activity prediction datasets to show that the DNN classifiers trained by the network configurations expressed by the final solutions of the PSO algorithm, employed to construct an ensemble model and individual classifier, outperform the random approach in terms of the generalization performance. Therefore, the proposed approach can be regarded an alternative tool for automatic network structure and parameter selection for deep neural networks. [ABSTRACT FROM AUTHOR]

In a modern small gasoline engine fuel injection system, the load of the engine is estimated based on the measurement of the manifold absolute pressure (MAP) sensor, which took place in the intake manifold. This paper present a more economical approach on estimating the MAP by using only the measurements of the throttle position and engine speed, resulting in lower implementation cost. The estimation was done via two-stage multilayer feed-forward neural network by combining Levenberg-Marquardt (LM) algorithm, Bayesian Regularization (BR) algorithm and Particle Swarm Optimization (PSO) algorithm. Based on the results found in 20 runs, the second variant of the hybrid algorithm yields a better network performance than the first variant of hybrid algorithm, LM, LM with BR and PSO by estimating the MAP closely to the simulated MAP values. By using a valid experimental training data, the estimator network that trained with the second variant of the hybrid algorithm showed the best performance among other algorithms when used in an actual retrofit fuel injection system (RFIS). The performance of the estimator was also validated in steady-state and transient condition by showing a closer MAP estimation to the actual value. [ABSTRACT FROM AUTHOR]

In this article, we propose a new data mining algorithm, by which one can both capture the non-linearity in data and also find the best subset model. To produce an enhanced subset of the original variables, a preferred selection method should have the potential of adding a supplementary level of regression analysis that would capture complex relationships in the data via mathematical transformation of the predictors and exploration of synergistic effects of combined variables. The method that we present here has the potential to produce an optimal subset of variables, rendering the overall process of model selection more efficient. This algorithm introduces interpretable parameters by transforming the original inputs and also a faithful fit to the data. The core objective of this paper is to introduce a new estimation technique for the classical least square regression framework. This new automatic variable transformation and model selection method could offer an optimal and stable model that minimizes the mean square error and variability, while combining all possible subset selection methodology with the inclusion variable transformations and interactions. Moreover, this method controls multicollinearity, leading to an optimal set of explanatory variables. [ABSTRACT FROM AUTHOR]

Gastric precancerous diseases (GPD) may deteriorate into early gastric cancer if misdiagnosed, so it is important to help doctors recognize GPD accurately and quickly. In this paper, we realize the classification of 3-class GPD, namely, polyp, erosion, and ulcer using convolutional neural networks (CNN) with a concise model called the Gastric Precancerous Disease Network (GPDNet). GPDNet introduces fire modules from SqueezeNet to reduce the model size and parameters about 10 times while improving speed for quick classification. To maintain classification accuracy with fewer parameters, we propose an innovative method called iterative reinforced learning (IRL). After training GPDNet from scratch, we apply IRL to fine-tune the parameters whose values are close to 0, and then we take the modified model as a pretrained model for the next training. The result shows that IRL can improve the accuracy about 9% after 6 iterations. The final classification accuracy of our GPDNet was 88.90%, which is promising for clinical GPD recognition. [ABSTRACT FROM AUTHOR]