Machine Learning Mechanisms Research Articles

Learning in Artificial Intelligence: Does Bloom's Taxonomy Apply?

From the early days of science and philosophy, humans have wondered exactly what it is that makes us intelligent beings. Plato felt that a good education should include instruction in music, gymnastics, and dialectic, and pondered and how we learned the things we did. Humans also been interested in building machines that mimic their abilities. The original motivation of building mechanical machines was to free us from mundane tasks and to allow fewer people to do particular difficult jobs. Today we build electronic machines that not only do our chores and to assist us physically, but also to help us intellectually.A side affect of building intelligent, electronic machines is that we come to know how humans work a little bit better, and we also learn to better appreciate our amazing ability to learn. There is a danger that we may take the results of artificial intelligence too seriously, and to extend them in ways that are not appropriate. The issues surrounding the mechanisms of artificial intelligence and machine learning may or may not apply to the way that people learn. They are simply the way that we imagine we may learn, a model that needs to be tested against reality. On the other hand, if it makes a computer appear intelligent then it is sufficient, then it is sufficient for our purposes whether it is the way people do it or not. The difference is in intent: are we attempting to automate some difficult process, or are we trying to determine how the human brain acquires new information?

Deep Filter Banks for Land-Use Scene Classification

Land-use (LU) scene classification is one of the most challenging tasks in the field of remote sensing (RS) image processing due to its high intraclass variability and low interclass distance. Motivated by the challenge posed by this problem, we propose a novel hybrid architecture, deep filter banks, combining multicolumn stacked denoising sparse autoencoder (SDSAE) and Fisher vector (FV) to automatically learn the representative and discriminative features in a hierarchical manner for LU scene classification. SDSAE kernels describe local patches and a robust global feature of the RS image is built through the FV pooling layer. Unlike previous handcrafted features, we use machine-learning mechanisms to optimize our proposed feature extractor so that it can learn more suitable internal features from the RS data, boosting the final performance. Our approach achieves superior performance compared with the state-of-the-art methods, obtaining average classification accuracies of 92.7% and 90.4%, respectively, on the UC Merced and RSSCN7 data sets.

Online Learners' Reading Ability Detection Based on Eye-Tracking Sensors.

The detection of university online learners’ reading ability is generally problematic and time-consuming. Thus the eye-tracking sensors have been employed in this study, to record temporal and spatial human eye movements. Learners’ pupils, blinks, fixation, saccade, and regression are recognized as primary indicators for detecting reading abilities. A computational model is established according to the empirical eye-tracking data, and applying the multi-feature regularization machine learning mechanism based on a Low-rank Constraint. The model presents good generalization ability with an error of only 4.9% when randomly running 100 times. It has obvious advantages in saving time and improving precision, with only 20 min of testing required for prediction of an individual learner’s reading ability.

Cyberdelegation and the Administrative State

This paper explores questions and trade-offs associated with delegating administrative agency decisions to computer algorithms, neural networks, and similar examples of “artificial intelligence," and offers the following preliminary observations to further discussion of the opportunities and risks. First, neither conventional expert systems nor neural networks (or other machine learning mechanisms) are in a position to resolve (without human intervention) context-specific debates about society’s goals for regulation or administrative adjudication – and these debates are often inherent in the implementation of statutes. Those goals must also inform whether we assign value to aspects of human cognition that contrast with what computers can (presently) accomplish, or what might be conventionally defined as rational in a decision-theoretic sense. Second, society must consider path-dependent consequences and associated cybersecurity risks that could arise from reliance on computers to make and support decisions. Such consequences include the erosion of individual and organizational knowledge over time. Third, it may prove difficult to limit the influence of computer programs even if they are meant to be mere decision support tools rather than the actual means of making a decision. Finally, heavy reliance on computer programs – particularly adaptive ones that modify themselves over time – may further complicate public deliberation about administrative decisions, because few if any observers will be entirely capable of understanding how a given decision was reached.

Stock Price Prediction Based on Financial Statements Using SVM

The support vector machine (SVM) is a fast, and reliable machine learning mechanism. In this paper, we evaluated the stock price predictability of SVM, which is a kind of fundamental analysis that predicts the stock price from corporate intrinsic value. Corporate financial statements are used as input into SVM. Based on the results, we predicted the rise or drop of the stock. In addition, we evaluated how long a given financial statement can be used to predict a stock’s price. Compared to the experts forecast, the results of SVM show good predictability. However, as times goes on, the predictability begins to drop. These predictions based on financial statements are excellent, but after a short period, the dissonance between financial statements and stock price can be offset by reasonable investors. These results support the efficient market hypothesis.

SVM 기반의 재무 정보를 이용한 주가 예측

Machine learning is a technique for training computers to be used in classification or forecasting. Among the various types, support vector machine (SVM) is a fast and reliable machine learning mechanism. In this paper, we evaluate the stock price predictability of SVM based on financial statements, through a fundamental analysis predicting the stock price from the corporate intrinsic values. Corporate financial statements were used as the input for SVM. Based on the results, the rise or drop of the stock was predicted. The SVM results were compared with the forecasts of experts, as well as other machine learning methods such as ANN, decision tree and AdaBoost. SVM showed good predictive power while requiring less execution time than the other machine learning schemes.

A comparative study of surface EMG classification by fuzzy relevance vector machine and fuzzy support vector machine

We present a multiclass fuzzy relevance vector machine (FRVM) learning mechanism and evaluate its performance to classify multiple hand motions using surface electromyographic (sEMG) signals. The relevance vector machine (RVM) is a sparse Bayesian kernel method which avoids some limitations of the support vector machine (SVM). However, RVM still suffers the difficulty of possible unclassifiable regions in multiclass problems. We propose two fuzzy membership function-based FRVM algorithms to solve such problems, based on experiments conducted on seven healthy subjects and two amputees with six hand motions. Two feature sets, namely, AR model coefficients and room mean square value (AR-RMS), and wavelet transform (WT) features, are extracted from the recorded sEMG signals. Fuzzy support vector machine (FSVM) analysis was also conducted for wide comparison in terms of accuracy, sparsity, training and testing time, as well as the effect of training sample sizes. FRVM yielded comparable classification accuracy with dramatically fewer support vectors in comparison with FSVM. Furthermore, the processing delay of FRVM was much less than that of FSVM, whilst training time of FSVM much faster than FRVM. The results indicate that FRVM classifier trained using sufficient samples can achieve comparable generalization capability as FSVM with significant sparsity in multi-channel sEMG classification, which is more suitable for sEMG-based real-time control applications.

Qualitative: Open Source Python Tool for Quality Estimation over Multiple Machine Translation Outputs

Abstract “Qualitative” is a python toolkit for ranking and selection of sentence-level output by different MT systems using Quality Estimation. The toolkit implements a basic pipeline for annotating the given sentences with black-box features. Consequently, it applies a machine learning mechanism in order to rank data based on models pre-trained on human preferences. The preprocessing pipeline includes support for language models, PCFG parsing, language checking tools and various other pre-processors and feature generators. The code follows the principles of object-oriented programming to allow modularity and extensibility. The tool can operate by processing both batch-files and single sentences. An XML-RPC interface is provided for hooking up with web-services and a graphical animated web-based interface demonstrates its potential on-line use.

Retrieval Model Based on Image Content Features

This paper presents fuzzy color histogram feature-based image retrieval method and texture spectrum fuzzy histogram feature analyzes the image database indexing techniques and the introduction of the experimental system for an improved method of fuzzy indexes. Algorithm reflects the underlying characteristics of high-level concepts and integration, relevance feedback and machine learning mechanism combining ideas. In this paper, the algorithm full processing power of computer systems, has a certain reference value and practical significance.

A Chaotic Complexity Measure for Cognitive Machine Classification of Cyber-Attacks on Computer Networks

Today's evolving cyber security threats demand new, modern, and cognitive computing approaches to network security systems. In the early years of the Internet, a simple packet inspection firewall was adequate to stop the then-contemporary attacks, such as Denial of Service (DoS), ports scans, and phishing. Since then, DoS has evolved to include Distributed Denial of Service (DDoS) attacks, especially against the Domain Name Service (DNS). DNS based DDoS amplification attacks cannot be stopped easily by traditional signature based detection mechanisms because the attack packets contain authentic data, and signature based detection systems look for specific attack-byte patterns. This paper proposes a chaos based complexity measure and a cognitive machine classification algorithm to detect DNS DDoS amplification attacks. In particular, this paper computes the Lyapunov exponent to measure the complexity of a flow of packets, and classifies the traffic as either normal or anomalous, based on the magnitude of the computed exponent. Preliminary results show the proposed chaotic measure achieved a detection (classification) accuracy of about 98%, which is greater than that of an Artificial Neural Network. Also, contrary to available supervised machine learning mechanisms, this technique does not require any offline training. This approach is capable of not only detecting offline threats, but has the potential of being applied over live traffic flows using DNS filters.

Teaching an Old Log New Tricks with Machine Learning.

To most people, the log file would not be considered an exciting area in technology today. However, these relatively benign, slowly growing data sources can drive large business transformations when combined with modern-day analytics. Accenture Technology Labs has built a new framework that helps to expand existing vendor solutions to create new methods of gaining insights from these benevolent information springs. This framework provides a systematic and effective machine-learning mechanism to understand, analyze, and visualize heterogeneous log files. These techniques enable an automated approach to analyzing log content in real time, learning relevant behaviors, and creating actionable insights applicable in traditionally reactive situations. Using this approach, companies can now tap into a wealth of knowledge residing in log file data that is currently being collected but underutilized because of its overwhelming variety and volume. By using log files as an important data input into the larger enterprise data supply chain, businesses have the opportunity to enhance their current operational log management solution and generate entirely new business insights-no longer limited to the realm of reactive IT management, but extending from proactive product improvement to defense from attacks. As we will discuss, this solution has immediate relevance in the telecommunications and security industries. However, the most forward-looking companies can take it even further. How? By thinking beyond the log file and applying the same machine-learning framework to other log file use cases (including logistics, social media, and consumer behavior) and any other transactional data source.

视觉感知结合学习的自然图像轮廓检测

A model for detecting contours in natural images is presented by combining the visual perceptual mechanisms of nonclassical receptive fields and machine learning. When the stimuli inside and outside the receptive field (RF) form a precise spatial configuration, the surround stimuli will enhance the response of the central stimulus, on the other hand, surround inhibition will reduce the responses to homogeneous elements. The two different visual mechanisms are used to enhance the well-organized structures and reduce the non-meaningful distractors engendering from texture fields, respectively. We approach the task of information combination as a supervised learning problem using the logistic regression model, where we will learn the combination rules from the ground truth data. Our experiments demonstrate that the model can dramatically suppress texture edges and reduce spurious contours, and meanwhile can enhance the responses to the elements with coherent spatial configurations and avoid ground-truth contours missed by the detector. Finally, we evaluate quantitatively our results with the Berkeley segmentation dataset and compare with related work. The proposed method not only provides a biologically feasible scheme for contour detection but also contributes to further understand visual mechanisms.

Energy-aware Activity Classification using Wearable Sensor Networks.

This paper presents implementation details, system characterization, and the performance of a wearable sensor network that was designed for human activity analysis. Specific machine learning mechanisms are implemented for recognizing a target set of activities with both out-of-body and on-body processing arrangements. Impacts of energy consumption by the on-body sensors are analyzed in terms of activity detection accuracy for out-of-body processing. Impacts of limited processing abilities for the on-body scenario are also characterized in terms of detection accuracy, by varying the background processing load in the sensor units. Impacts of varying number of sensors in terms of activity classification accuracy are also evaluated. Through a rigorous systems study, it is shown that an efficient human activity analytics system can be designed and operated even under energy and processing constraints of tiny on-body wearable sensors.

Web-Based Newborn Screening System for Metabolic Diseases: Machine Learning Versus Clinicians

BackgroundA hospital information system (HIS) that integrates screening data and interpretation of the data is routinely requested by hospitals and parents. However, the accuracy of disease classification may be low because of the disease characteristics and the analytes used for classification.ObjectiveThe objective of this study is to describe a system that enhanced the neonatal screening system of the Newborn Screening Center at the National Taiwan University Hospital. The system was designed and deployed according to a service-oriented architecture (SOA) framework under the Web services .NET environment. The system consists of sample collection, testing, diagnosis, evaluation, treatment, and follow-up services among collaborating hospitals. To improve the accuracy of newborn screening, machine learning and optimal feature selection mechanisms were investigated for screening newborns for inborn errors of metabolism.MethodsThe framework of the Newborn Screening Hospital Information System (NSHIS) used the embedded Health Level Seven (HL7) standards for data exchanges among heterogeneous platforms integrated by Web services in the C# language. In this study, machine learning classification was used to predict phenylketonuria (PKU), hypermethioninemia, and 3-methylcrotonyl-CoA-carboxylase (3-MCC) deficiency. The classification methods used 347,312 newborn dried blood samples collected at the Center between 2006 and 2011. Of these, 220 newborns had values over the diagnostic cutoffs (positive cases) and 1557 had values that were over the screening cutoffs but did not meet the diagnostic cutoffs (suspected cases). The original 35 analytes and the manifested features were ranked based on F score, then combinations of the top 20 ranked features were selected as input features to support vector machine (SVM) classifiers to obtain optimal feature sets. These feature sets were tested using 5-fold cross-validation and optimal models were generated. The datasets collected in year 2011 were used as predicting cases.ResultsThe feature selection strategies were implemented and the optimal markers for PKU, hypermethioninemia, and 3-MCC deficiency were obtained. The results of the machine learning approach were compared with the cutoff scheme. The number of the false positive cases were reduced from 21 to 2 for PKU, from 30 to 10 for hypermethioninemia, and 209 to 46 for 3-MCC deficiency.ConclusionsThis SOA Web service–based newborn screening system can accelerate screening procedures effectively and efficiently. An SVM learning methodology for PKU, hypermethioninemia, and 3-MCC deficiency metabolic diseases classification, including optimal feature selection strategies, is presented. By adopting the results of this study, the number of suspected cases could be reduced dramatically.

A survey of multi-view machine learning

Multi-view learning or learning with multiple distinct feature sets is a rapidly growing direction in machine learning with well theoretical underpinnings and great practical success. This paper reviews theories developed to understand the properties and behaviors of multi-view learning and gives a taxonomy of approaches according to the machine learning mechanisms involved and the fashions in which multiple views are exploited. This survey aims to provide an insightful organization of current developments in the field of multi-view learning, identify their limitations, and give suggestions for further research. One feature of this survey is that we attempt to point out specific open problems which can hopefully be useful to promote the research of multi-view machine learning.

Automated quantification of spikes

Methods for rapid and objective quantification of interictal spikes in raw, unprocessed electroencephalogram (EEG) samples are scarce. We evaluated the accuracy of a tailored automated spike quantification algorithm. The automated quantification was compared with the quantification by two board-certified clinical neurophysiologists (gold-standard) in five steps: 1) accuracy in a single EEG channel (5 EEG samples), 2) accuracy in multiple EEG channels and across different stages of the sleep-wake cycles (75 EEG samples), 3) capacity to detect lateralization of spikes (6 EEG samples), 4) accuracy after application of a machine-learning mechanism (11 EEG samples), and 5) accuracy during wakefulness only (8 EEG samples). Our method was accurate during all stages of the sleep-wake cycle and improved after the application of the machine-learning mechanism. Spikes were correctly lateralized in all cases. Our automated method was accurate in quantifying and detecting the lateralization of interictal spikes in raw unprocessed EEG samples.

Machine learning mechanisms to diagnose computer virus.

Machine learning mechanisms to diagnose computer virus.

Self-adaptive large neighborhood search algorithm for parallel machine scheduling problems

A self-adaptive large neighborhood search method for scheduling n jobs on m non-identical parallel machines with multiple time windows is presented. The problems' another feature lies in oversubscription, namely not all jobs can be scheduled within specified scheduling horizons due to the limited machine capacity. The objective is thus to maximize the overall profits of processed jobs while respecting machine constraints. A first-in-first-out heuristic is applied to find an initial solution, and then a large neighborhood search procedure is employed to relax and re-optimize cumbersome solutions. A machine learning mechanism is also introduced to converge on the most efficient neighborhoods for the problem. Extensive computational results are presented based on data from an application involving the daily observation scheduling of a fleet of earth observing satellites. The method rapidly solves most problem instances to optimal or near optimal and shows a robust performance in sensitive analysis.

Discriminating Outer Membrane Proteins with Fuzzy K-Nearest Neighbor Algorithms Based on the General Form of Chou’s PseAAC

Outer membrane proteins (OMPs) play important roles in cell biology. In addition, OMPs are targeted by multiple drugs. The identification of OMPs from genomic sequences and successful prediction of their secondary and tertiary structures is a challenging task due to short membrane-spanning regions with high variation in properties. Therefore, an effective and accurate silico method for discrimination of OMPs from their primary sequences is needed. In this paper, we have analyzed the performance of various machine learning mechanisms for discriminating OMPs such as: Genetic Programming, K-nearest Neighbor, and Fuzzy K-nearest Neighbor (Fuzzy K-NN) in conjunction with discrete methods such as: Amino acid composition, Amphiphilic Pseudo amino acid composition, Split amino acid composition (SAAC), and hybrid versions of these methods. The performance of the classifiers is evaluated by two datasets using 5-fold crossvalidation. After the simulation, we have observed that Fuzzy K-NN using SAAC based-features makes it quite effective in discriminating OMPs. Fuzzy K-NN achieves the highest success rates of 99.00% accuracy for discriminating OMPs from non-OMPs and 98.77% and 98.28% accuracies from α-helix membrane and globular proteins, respectively on dataset1. While on dataset2, Fuzzy K-NN achieves 99.55%, 99.90%, and 99.81% accuracies for discriminating OMPs from non- OMPs, α-helix membrane, and globular proteins, respectively. It is observed that the classification performance of our proposed method is satisfactory and is better than the existing methods. Thus, it might be an effective tool for high throughput innovation of OMPs.

Expectation of trading agent behaviour in negotiation of electronic marketplace

Electronic Commerce has been a very significant commercial phenomenon in recent years, and autonomous agents are widely adopted by business or individuals in electronic marketplaces to fulfill time consuming tasks in trading. Agent negotiation mechanisms are usually applied between conflicted agents in order to reach a mutually beneficial agreement. Prediction of trading agents' strategies and behaviours in negotiation is a very significant research topic in agent negotiation. By employing the prediction results on opponents' possible strategies and behaviours during a negotiation, trading agents can plan and perform corresponding strategies in order to maximize their own profits. Significant achievements have been made on this topic. However, most existing approaches are based on machine learning mechanisms, which may fail to capture opponents' behaviours in open and dynamic electronic marketplaces. In this paper, two agent behaviour expectation approaches are introduced to help trading agents to capture opponents' potential behaviours during a negotiation in complex e-marketplaces. i The regression analysis approach focuses on illustrating the main trends of opponents' trading behaviours; ii the vector analysis approach pays more attention to identifying opponents' detailed negotiation strategies. The experimental results show the efficiency and efficacy of the two proposed approaches in open and dynamic negotiation environments.