Chan-Yun Yang

Margin calibration in SVM class-imbalanced learning

Imbalanced dataset learning is an important practical issue in machine learning, even in support vector machines (SVMs). In this study, a well known reference model for solving the problem proposed by Veropoulos et al., is first studied. From the aspect of loss function, the reference cost sensitive prototype is identified as a penalty-regularized model. Intuitively, the loss function can change not only the penalty but also the margin to recover the biased decision boundary. This study focuses mainly on the effect from the margin and then extends the model to a more general modification. As proposed in the prototype, the modification first adopts an inversed proportional regularized penalty to re-weight the imbalanced classes. In addition to the penalty regularization, the modification then employs a margin compensation to lead the margin to be lopsided, which enables the decision boundary drift. Two regularization factors, the penalty and margin, are hence suggested for achieving an unbiased classification. The margin compensation, associating with the penalty regularization, is here utilized to calibrate and refine the biased decision boundary to further reduce the bias. With the area under the receiver operating characteristic curve (AuROC) for examining the performance, the modification shows relative higher scores than the reference model, even though the optimal performance is achieved by the reference model. Some useful characteristics found empirically are also included, which may be convenient for the future applications. All the theoretical descriptions and experimental validations show the proposed models potential to compete for highly unbiased accuracy in a complex imbalanced dataset.

Grey decision-making for a billiard robot

Billiard is one of the most complex games to play in the real world. A player needs to visualize the situation between balls and pockets and to score the ball into the designate pocket by his/her own experience. A billiard robot is developed to imitate the behavior of human beings to play billiard. There are machine vision, decision-making, control and actuating subsystems in the experiment setup. The objective of this paper is to design a decision algorithm for a billiard robot by using grey theory. The results indicate that the decision algorithm work very well in both the simulation and experiment.

Classification of rotifers with machine vision by shape moment invariants

An automated system for the identification of rotifers under a microscope with machine vision by shape analysis has been developed, which tends to be substituted for human appraisal. A suitable image recognition algorithm was proposed and the results were discussed in detail. In this study, rotifers were classified into the exact types despite the debris, which appeared from sludge in the degraded water or from rotifer carcasses. Two stages of a discrimination model based on shape analysis were built: one was to separate debris from rotifers, and the other was to classify rotifers into three groups. A set of shape descriptors, including geometry and moment features, was extracted from the images. The set of shape descriptors had to satisfy the RST (rotation, scaling, and translation) invariance. Shape analysis was proved to be an effective approach since the classification accuracy was approx. 92%. The results from different classification approaches were also compared. The machine vision system with shape analysis and the 2-stage discrimination model had a greater effect on the reduction of manpower requirement for the classification of rotifers.

Imbalanced SVM Learning with Margin Compensation

The paper surveys the previous solutions and proposes further a new solution based on the cost-sensitive learning for solving the imbalanced dataset learning problem in the support vector machines. The general idea of cost-sensitive approach is to adopt an inverse proportional penalization scheme for dealing with the problem and forms a penalty regularized model. In the paper, additional margin compensation is further included to achieve a more accurate solution. As known, the margin plays an important role in drawing the decision boundary. It motivates the study to produce imbalanced margin between the classes which enables the decision boundary shift. The imbalanced margin is hence allowed to recompense the overwhelmed class as margin compensation. Incorporating with the penalty regularization, the margin compensation is capable to calibrate moderately the decision boundary and can be utilized to refine the bias boundary. The effect decreases the need of high penalty on the minority class and prevents the classification from the risk of overfitting. Experimental results show a promising potential in future applications.

Approximation of algebraic and trigonometric polynomials by feedforward neural networks

This paper presents a function approximation to a general class of polynomials by using one-hidden-layer feedforward neural networks(FNNs). Both the approximations of algebraic polynomial and trigonometric polynomial functions are discussed in details. For algebraic polynomial functions, an one-hidden-layer FNN with chosen number of hidden-layer nodes and corresponding weights is established by a constructive method to approximate the polynomials to a remarkable high degree of accuracy. For trigonometric functions, an upper bound of approximation is therefore derived by the constructive FNNs. In addition, algorithmic examples are also included to confirm the accuracy performance of the constructive FNNs method. The results show that it improves efficiently the approximations of both algebraic polynomials and trigonometric polynomials. Consequently, the work is really of both theoretical and practical significance in constructing a one-hidden-layer FNNs for approximating the class of polynomials. The work also paves potentially the way for extending the neural networks to approximate a general class of complicated functions both in theory and practice.

Support Vector Classifier with a Fuzzy-Value Class Label

The purpose of this paper is to introduce a concept of fuzzy class memberships to the samples of training set in the support vector classifier. The inclusion of fuzzy values contributed a set of dynamic Lagrangian constraints, which setups a more specific space for searching the optimum, and conducted a more accurate classification performance. The developed model stepped into the sub-structure of the classifier, and involved the complex micro-interactions among the training samples to form a more precise separating hyperplane by fuzzy membership. The micro-interactions also altered the hyperplane and its corresponding margin, and achieved the deep-reaching classification accuracy around the sub-optimal region.

Development of Robotic Grasping Gripper Based on Smart Fuzzy Controller

Abstract An alternative robotic grasping gripper including a vision system, machine fingers, pressure modules, and smart fuzzy grasping controller is designed and implemented in this paper. To avoid the redundant computation of inverse kinematics, the relative coordinates are adopted in the proposed architecture. To identify the stiffness and shape of different grasping objects, a smart fuzzy grasping controller is embedded into the recognition process first. According to the identifying results, the membership functions of the smart fuzzy grasping controller are precisely tuned to generate the joint angles of the servo motors online. The effectiveness is verified by some experimental results, and the proposed architectures are implemented in the home-made robotic grasping gripper in laboratory.

Internet of Things for human — Pet interaction

Nowadays, factors such as aging from one side and declining birthrate on the other side in addition to modern workforce and change in social structure has strong impact on relationships in the society. As one result of the above factors many people prefer to own a pet as a companion. However, due to the busy work schedule for most of the pet owners and not being able to share the caring task with others, paying full time attention to the pet is almost impossible. To solve the mentioned problem, we aimed to develop intelligent and interactive system to bridge the gap between the pet and the pet owner. This research is based on the concept of the Internet of Things by using Linux operating system and Raspberry Pi board as development platform. The proposed system provides a smart phone application to the pet owner that can make real-time communication with the pet using internet. As an input module the owner can use the vision module of the system to have a view of his/her pet and as an output module can control the servo motors of the wearable device on the pet for feeding. In future, we hope this system can be improved and used by pet owners for real-time interaction with their pets from distance where they can be in the same location physically.

Adaptation of mobile robots to intelligent vehicles

Nowadays, robots are used to perform tasks that require great levels of precision or are simply repetitive. Mobile robots are able to move their position from one place to another in unstructured or structured environments to do their task with using sensor for their navigation. In this paper, we present some of current technology in mobile robots which can be implemented to intelligent vehicles based on their architecture that uses a sequence of three steps: Sense, Plan, and Act (SPA). For sensing, several sensors such as camera and proximity sensors are common modules for a mobile robot which can be also employed in an intelligent vehicle. Advanced technologies for planning in mobile robots can be employed to intelligent vehicles in order to organize and plan more intelligent behaviors. Furthermore trends in actuation of the mobile robot can be adapted to vehicles to achieve novel architecture in future cars.

Associating k NN and SVM for higher classification accuracy

The paper proposed a hybrid two-stage method of support vector machines (SVM) to increase its performance in classification accuracy. In this model, a filtering stage of the k nearest neighbor (kNN) rule was employed to collect information from training observations and re-evaluate balance weights for the observations based on their influences. The balance weights changed the policy of the discrete class label. A novel idea of real-valued class labels for transferring the balance weights was therefore proposed. Embedded in the class label, the weights given as the penalties of the uncertain outliers in the classification were considered in the quadratic programming of SVM, and produced a different hyperplane with higher accuracy. The adoption of kNN rule in the filtering stage has the advantage to distinguish the uncertain outliers in an independent way. The results showed that the classification accuracy of the hybrid model was higher than that of the classical SVM.