Una M. Kelly, Luuk Spreeuwers and Raymond Veldhuis, Data Management and Biometrics Group, University of Twente, The Netherlands
State-of-the-art face recognition systems (FRS) are vulnerable to morphing attacks, in which two photos of different people are merged in such a way that the resulting photo resembles both people. Such a photo could be used to apply for a passport, allowing both people to travel with the same identity document. Research has so far focussed on developing morphing detection methods. We suggest that it might instead be worthwhile to make face recognition systems themselves more robust to morphing attacks. We show that deep-learning-based face recognition can be improved simply by treating morphed images just like real images during training but also that, for significant improvements, more work is needed. Furthermore, we test the performance of our FRS on morphs of a type not seen during training. This addresses the problem of overfitting to the type of morphs used during training, which is often overlooked in current research.
Biometrics, Morphing Attack Detection, Face Recognition, Vulnerability of Biometric Systems.
Yuh-Jen Chen, Department of Accounting and Information Systems, National Kaohsiung University of Science and Technology, Kaohsiung, Taiwan, ROC
Financial forecasts are regarded as vital financial information for most enterprises. They not only project the financial performance of an enterprise in a future operating period but also assist internal managers with operations, investment, and financing decision-making and external investors and creditors with understanding the operating performance of the enterprise. However, a financial forecast of an enterprise must be comprehensive to rule out unreasonable assumptions arising from local forecasts. Therefore, finding ways to assist enterprises with producing accurate comprehensive financial forecasts has become a critical issue in research on financial management. In consideration of the financial indicators of financial structure, solvency, operating ability, profitability, and cash flow as well as the non-financial indicators of firm size and corporate governance, the algorithms of multivariate adaptive regression splines (MARS) and queen genetic algorithm-support vector regression (QGA-SVR) are used in this study to create a comprehensive financial forecast of operating revenue, earnings per share, free cash flow, and net working capital to help enterprises forecast their future financial situation and offer investors and creditors a reference for investment decision-making. This study’s objectives are achieved through the following steps: (i) establishment of feature indicators for financial forecasting, (ii) development of a financial forecasting method, and (iii) demonstration of the proposed method and comparison with existing methods.
Financial Forecasting, Multivariate Adaptive Regression Splines (MARS), Queen Genetic Algorithm (QGA), Support Vector Regression (SVR).
Saranya M1 and Geetha T V2, 1Computer Science and Engineering, CEG, Anna University, India, 2Senior Professor, Computer Science and Engineering, CEG, Anna University, India
Now-a-days people around the world are infected by many new diseases. Developing or discovering a new drug for the newly discovered disease is an expensive and time consuming process and these could be eliminated if already existing resources could be used. For identifying the candidates from available drugs we need to perform text mining of a large-scale literature repository to extract the relation between the chemical, target and disease. Computational approaches for identifying the relationships between the entities in biomedical domain are appearing as an active area of research for drug discovery as it requires more man power. Currently, the computational approaches for extracting the biomedical relations such as drug-gene and gene-disease relationships are limited as the construction of drug-gene and gene-disease association from the unstructured biomedical documents is very hard. In this work, we propose pattern based bootstrapping method which is a semi-supervised learning algorithm to extract the direct relations between drug, gene and disease from the biomedical documents. These direct relationships are used to infer indirect relationships between entities such as drug and disease. Now these indirect relationships are used to determine the new candidates for drug repositioning which in turn will reduce the time and the patient’s risk.
Text mining, drug discovery, drug repositioning, bootstrapping, machine learning.
Ram chandra Pal, Dr A.P.J.Abdul Kalam University Indore, India
Social media is one of the biggest forums to express opinions. Sentiment analysis is the procedure by which information is extracted from the opinions, appraisals and emotions of people in regards to entities, events and their attributes. Sentiment analysis is also known as opinion mining. Opinion Mining is to analyses and classifies the user generated data like reviews, blogs, comments, articles etc. The main objective of Opinion mining is Sentiment Classification i.e. to classify the opinion into positive or negative classes. A earlier work is based on star rating of user data, most of the reviews are written in text format. The reviews are in text format which is difficult for computer system to understand. In recent internet applications, which have focused on detecting the polarity of the text, our text classifier helps users distinguish between positive and negative reviews thus assisting the user with opinion Extraction. This could be very useful for web applications like twitter, where the user has to face large chunks of raw data. To classify opinion an unsupervised lexicon technique is used for sentiment classification. There are so many user generated opinions on the web for a product; it may be difficult to know how many opinions are positive or negative. It makes tough to take decision about product purchasing. a sentence level opinion Extraction is used and it is done by counting based approach which compare the opinion by count method. All customer reviews of product need to summarize; we do not summarize the reviews by selecting or rewriting a subset of the original sentence from the reviews.
Machine Learning Algorithms, Opinion Extraction, Web Text, customer reviews.
Zhuang Liu and Yuanping Zhu, School of Computer and Information Engineering, Tianjin Normal University, Tianjin, China
This paper studies the license plate recognition problem under the complex background and the license plate tilt. Existing methods cannot solve these problems well. This paper proposes an end-to-end correction network based on deep learning. The model contains three parts: correction network, residual module and sequence module, which are responsible for distortion of license plate correction, image feature extraction and license plate character recognition. In the experiments, we studied the effects of complex backgrounds such as light, rain and snow, and the inclination and distortion of license plates on the accuracy of license plate recognition. The experimental part of this article uses the Chinese Academy of Sciences CCPD dataset, which covers almost all license plate data in natural scenes. The experimental results show that compared with the existing license plate recognition algorithm, the algorithm in this paper achieves an accuracy improvement on the test project, and it averages about 5% in complex scenarios.
Correction Network, Convolutional Neural Network, License Plate Recognition, Smart Transportation.
Jianyong XUE1, Olivier L. Georgeon1,2 and SalimaHassas1, 1LIRIS CNRS UMR5205, Université Claude Bernard Lyon 1, Lyon, France, 2LBG UMRS 449, Université Catholique de Lyon, Lyon, France
During the initial phase of cognitive development, infants exhibit amazing abilities to generate novel behaviours in unfamiliar situations, and explore actively to learn the best while lacking extrinsic rewards from the environment. These abilities set them apart from even the most advanced autonomous robots. This work seeks to contribute to understand and replicate some of these abilities. We propose the Bottom-up hiErarchical sequential Learning algorithm with Constructivist pAradigm (BEL-CA) to design agents capable of learning autonomously and continuously through interaction. The algorithm implements no assumption about the semantics of input and output data, nor relies upon a model of the world given a priori in the form of a set of states and transitions as well. Besides, we propose a toolkit to analyse the learning process at run time called GAIT (Generating and Analysing Interaction Traces). We use GAIT to report and explain the detailed learning process and the structured behaviours that the agent learns on each decision making. We report an experiment in which the agent learned to successfully interact with its environment and to avoid unfavourable interactions using regularities discovered through interaction.
cognitive development, constructivist learning, hierarchical sequential learning, self-adaptation.
A. V. H Sai Prasad1*, Dr. G. V. S. Rajkumar2, 1Research Scholar, Department of Computer Science and Engineering, GITAM Institute of Technology, 2Professor, Department of Computer science and Engineering, GITAM Institute of Technology, GITAM (Deemed to be University), Visakhapatnam - 530045, India
In the internet, a number of services have become flexible and cost-effective because of cloud computing. Security is the major hitch in cloud computing and many researchers have studied and discussed the problems relating to this issue. Various techniques are requiringensuring the integrity of data which is the integral part of cloud storage adoption. Five different trust attributes are collected from third party and its trust model in this work and integrity of data are assured through the servers.For optimal scheduling Ant Lion Optimizer (ALO) algorithm is used which is proposed and contrasted with Particle Swarm Optimization (PSO).
Cloud computing, data integrity, third party trust model, Particle Swarm Optimization (PSO) and Ant Lion Optimizer (ALO) Algorithm.
Ruben Ventura, Independent Security Researcher
This paper presents new and evolved methods to perform Blind SQL Injection attacks. These are much faster than the current publicly available tools and techniques due to various reasons. Implementing these methods within carefully crafted code has resulted in the development of the fastest tools in the world to extract information from a database through Blind SQL Injection vulnerabilities. The nature of such attack vectors will be explained in this paper, including all of their intrinsic details.
Web Application Security, Blind SQL Injection, Attack Optimization, New Exploitation Methods.
Björn Friedrich, Enno-Edzard Steen, Sebastian Fudickar and Andreas Hein, Department of Health Services Research, Carl von Ossietzky University, Oldenburg, Germany
A continuous monitoring of the physical strength and mobility of elderly people is important for maintaining their health and treating diseases at an early stage. However, frequent screenings by physicians are exceeding the logistic capacities. An alternate approach is the automatic and unobtrusive collection of functional measures by ambient sensors. In the current publication, we show the correlation among data of ambient motion sensors and the well-established mobility assessment Short-Physical-Performance-Battery and Tinetti. We use the average number of motion sensor events for correlation with the assessment scores. The evaluation on a real-world dataset shows a moderate to strong correlation with the scores of standardised geriatrics physical assessments.
ubiquitous computing, biomedical informatics, health, correlation, piecewise linear approximation.
Cao Xiaopeng and Qu Hongyan, School of Computer Science and Technology, Xi’an University of Posts and Telecommunications, Xi’an, China
The massive network traffic and high-dimensional features affect detection performance. In order to improve the efficiency and performance of detection, whale optimization sparse autoencoder model (WO-SAE)isproposed. Firstly, sparse autoencoder performs unsupervised training on high-dimensional raw data and extracts low- dimensional features of network traffic. Secondly, the key parameters of sparse autoencoder are optimized automatically by whale optimization algorithm to achieve better feature extraction ability. Finally, gated recurrent unit is used to classify the time series data. The experimental results show that the proposed model is superior to existing detection algorithms in accuracy, precision, and recall. And the accuracypresents 98.69%. WO-SAE model is a novel approach that reduces the user’s reliance on deep learning expertise.
Traffic anomaly detection, Feature extraction, Sparse autoencoder, Whale optimization algorithm.
Cao Xiaopeng and Shi Linkai, School of Computer Science and Technology, Xi’an University of Posts and Telecommunications, Xi’an, China
The practical Byzantine fault-tolerant algorithm does not add nodes dynamically. It is limited in practical application. In order to add nodes dynamically, Dynamic Practical Byzantine Fault Tolerance Algorithm (DPBFT) was proposed. Firstly, a new node sends request information to other nodes in the network. The nodes in the network decide their identities and requests. Then the nodes in the network reverse connect to the new node and send block information of the current network, the new node updates information. Finally, the new node participates in the next round of consensus, changes the view and selects the master node. This paper abstracts the decision of nodes into the undirected connected graph. The final consistency of the graph is used to prove that the proposed algorithm can adapt to the network dynamically.Compared with the PBFT algorithm, DPBFT has better fault tolerance and lower network bandwidth.
Practical Byzantine Fault Tolerance, Blockchain, Consensus Algorithm, Consistency Analysis.
Morio Yamauchi1, Kazuhisa Naakano2, Yoshiya Tanaka2 And Keiichi Horio1, 1Department of Human Intelligence Systems, Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology, Kitakyushu, Japan, 2The First Department of International Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan
In this article, we implemented a regression model and conducted experiments for predicting disease activity using data from 1929 rheumatoid arthritis patients to assist in the selection of biologics for rheumatoid arthritis. On modelling, the missing variables in the data were completed by three different methods, mean value, self-organizing map and random value. Experimental results showed that the prediction error of the regression model was large regardless of the missing completion method, making it difficult to predict the prognosis of rheumatoid arthritis patients.
Rheumatoid Arthritis, Gaussian Process Regression, Self-Organizing Map.