Real-life Application Domains Research Articles

DIESEL: A novel deep learning-based tool for SpMV computations and solving sparse linear equation systems

Sparse linear algebra is central to many areas of engineering, science, and business. The community has done considerable work on proposing new methods for sparse matrix-vector multiplication (SpMV) computations and iterative sparse solvers on graphical processing units (GPUs). Due to vast variations in matrix features, no single method performs well across all sparse matrices. A few tools on automatic prediction of best-performing SpMV kernels have emerged recently and require many more efforts to fully utilize their potential. The utilization of a GPU by the existing SpMV kernels is far from its full capacity. Moreover, the development and performance analysis of SpMV techniques on GPUs have not been studied in sufficient depth. This paper proposes DIESEL, a deep learning-based tool that predicts and executes the best performing SpMV kernel for a given matrix using a feature set carefully devised by us through rigorous empirical and mathematical instruments. The dataset comprises 1056 matrices from 26 different real-life application domains including computational fluid dynamics, materials, electromagnetics, economics, and more. We propose a range of new metrics and methods for performance analysis, visualization, and comparison of SpMV tools. DIESEL provides better performance with its accuracy \(88.2\%\), workload accuracy \(91.96\%\), and average relative loss \(4.4\%\), compared to \(85.9\%\), \(85.31\%\), and \(7.65\%\) by the next best performing artificial intelligence (AI)-based SpMV tool. The extensive results and analyses presented in this paper provide several key insights into the performance of the SpMV tools and how these relate to the matrix datasets and the performance metrics, allowing the community to further improve and compare basic and AI-based SpMV tools in the future.

Gentle Versus Strong Touch Classification: Preliminary Results, Challenges, and Potentials.

Touch plays a crucial role in humans’ nonverbal social and affective communication. It then comes as no surprise to observe a considerable effort that has been placed on devising methodologies for automated touch classification. For instance, such an ability allows for the use of smart touch sensors in such real-life application domains as socially-assistive robots and embodied telecommunication. In fact, touch classification literature represents an undeniably progressive result. However, these results are limited in two important ways. First, they are mostly based on overall (i.e., average) accuracy of different classifiers. As a result, they fall short in providing an insight on performance of these approaches as per different types of touch. Second, they do not consider the same type of touch with different level of strength (e.g., gentle versus strong touch). This is certainly an important factor that deserves investigating since the intensity of a touch can utterly transform its meaning (e.g., from an affectionate gesture to a sign of punishment). The current study provides a preliminary investigation of these shortcomings by considering the accuracy of a number of classifiers for both, within- (i.e., same type of touch with differing strengths) and between-touch (i.e., different types of touch) classifications. Our results help verify the strength and shortcoming of different machine learning algorithms for touch classification. They also highlight some of the challenges whose solution concepts can pave the path for integration of touch sensors in such application domains as human–robot interaction (HRI).

A Preliminary Study on Learning Challenges in Machine Learning-based Flight Delay Prediction

Machine learning based flight delay prediction is one of the numerous real-life application domains where the problem of imbalance in class distribution is reported to affect the performance of learning algorithms. However, the fact that learning algorithms have been reported to perform well on some class imbalance problems posits the possibility of other contributing factors. In this study, we visually explore air traffic data after dimensionality reduction with t-Distributed Stochastic Neighbour Embedding. Our initial findings suggest a high degree of overlapping between the delayed and on-time class instances which can be a greater problem for learning algorithms than class imbalance.

Radial-Based oversampling for noisy imbalanced data classification

Imbalanced data classification remains a focus of intense research, mostly due to the prevalence of data imbalance in various real-life application domains. A disproportion among objects from different classes may significantly affect the performance of standard classification models. The first problem is the high imbalance ratios that pose a serious learning difficulty and require usage of dedicated methods, capable of alleviating this issue. The second important problem which may appear is noise, which may be accompanying the training data and causing strong deterioration of the classifier performance or increase the time required for its training. Therefore, the desirable classification model should be robust to both skewed data distributions and noise. One of the most popular approaches for handling imbalanced data is oversampling of the minority objects in their neighborhood. In this work we will criticize this approach and propose a novel strategy for dealing with imbalanced data, with particular focus on the noise presence. We propose Radial-Based Oversampling (RBO) method, which can find regions in which the synthetic objects from minority class should be generated on the basis of the imbalance distribution estimation with radial basis functions. Results of experiments, carried out on a representative set of benchmark datasets, confirm that the proposed guided synthetic oversampling algorithm offers an interesting alternative to popular state-of-the-art solutions for imbalanced data preprocessing.

Visual Analysis of Collective Anomalies Using Faceted High-Order Correlation Graphs

Successfully detecting, analyzing, and reasoning about collective anomalies is important for many real-life application domains (e.g., intrusion detection, fraud analysis, software security). The primary challenges to achieving this goal include the overwhelming number of low-risk events and their multimodal relationships, the diversity of collective anomalies by various data and anomaly types, and the difficulty in incorporating the domain knowledge of experts. In this paper, we propose the novel concept of the faceted High-Order Correlation Graph (HOCG). Compared with previous, low-order correlation graphs, HOCG achieves better user interactivity, computational scalability, and domain generality through synthesizing heterogeneous types of objects, their anomalies, and the multimodal relationships, all in a single graph. We design elaborate visual metaphors, interaction models, and the coordinated multiple view based interface to allow users to fully unleash the visual analytics power of the HOCG. We conduct case studies for three application domains and collect feedback from domain experts who apply our method to these scenarios. The results demonstrate the effectiveness of the HOCG in the overview of point anomalies, the detection of collective anomalies, and the reasoning process of root cause analyses.

CUBu: Universal Real-Time Bundling for Large Graphs.

Visualizing very large graphs by edge bundling is a promising method, yet subject to several challenges: speed, clutter, level-of-detail, and parameter control. We present CUBu, a framework that addresses the above problems in an integrated way. Fully GPU-based, CUBu bundles graphs of up to a million edges at interactive framerates, being over 50 times faster than comparable state-of-the-art methods, and has a simple and intuitive control of bundling parameters. CUBu extends and unifies existing bundling techniques, offering ways to control bundle shapes, separate bundles by edge direction, and shade bundles to create a level-of-detail visualization that shows both the graph core structure and its details. We demonstrate CUBu on several large graphs extracted from real-life application domains.

An emerging hybrid mechanism for information disclosure forecasting

Corporate governance mechanisms ensure that investors get a fair return on their investment. A well-established governance mechanism reduces the information asymmetry and agency cost between a firm’s management and stakeholders, but decision makers find it difficult to assess the corporate governance status of publicly-listed firms before the annual official announcement the following year. This study proposes a hybrid ensemble learning forecasting mechanism (HELM), whose single-component candidates from the extreme learning machine (ELM) algorithm with dissimilar ensemble strategies (that is, data diversity, parameter diversity, kernel diversity, and pre-processing diversity) form one initial dataset. We implement locally linear embedding into the proposed mechanism to handle the dimensionality task and then utilize the weighted voting taken from the base components’ cross-validation performance on a training dataset as the integration mechanism. Experimental results show that the proposed HELM significantly outperforms the other classifiers, but its superior performance under many real-life application domains comes with a critical drawback: it is incapable of providing an explanation for the underlying reasoning mechanisms. Thus, this study advances the utilized rough set theory with its explanation capability to extract the inherent knowledge from the ensemble mechanism (HELM). The informative rules can be used as a guideline for decision makers to make a reliable judgment under turbulent financial markets.

A PARAMETER-BASED COMBINED CLASSIFIER FOR INVARIANT FACIAL EXPRESSION AND GENDER RECOGNITION

In this paper an learning vector quantization network architecture based on varying parameters and eliminating blind neurons28 is developed that learns the correlation of gender patterns and recognizes facial expressions of human faces. The network is developed to classify the 1 — neutral, 2 — smile or happiness, 3 — anger and 4 — scream or fear expressions. A peak accuracy rate of (a) 100% on the expression recognition task and (b) 100% on the gender recognition task for random training and test samples is achieved. The computer execution time for the recognition is about 0.004 seconds per face image for the latex expression classification task and 0.02 for the gender recognition task on an IBM PC. It is demonstrated that the proposed architecture is capable of achieving better recognition results with both good generalization performance and a fast training-time on a variety of test problems. The developed system showed potential for real life application domains.