International Journal on Artificial Intelligence Tools

A transfer learning environment is characterized by not having sufficient labeled training data from the domain of interest (target domain) to build a high-performing machine learner. Transfer learning algorithms use labeled data from an alternate domain (source domain), that is similar to the target domain, to build high-performing learners. The design of a transfer learning algorithm is typically comprised of a domain adaptation step following by a learning step. The domain adaptation step attempts to align the distribution differences between the source domain and the target domain. Then, the aligned data from the domain adaptation step is used in the learning step, which is typically implemented with a traditional machine learning algorithm. Our research studies the impact of the learning step on the performance of various transfer learning algorithms. In our experiment, we use five unique domain adaptation methods coupled with seven different traditional machine learning methods to create 35 different transfer learning algorithms. We perform comparative performance analyses of the 35 transfer learning algorithms, along with the seven stand-alone traditional machine learning methods. This research will aid machine learning practitioners in the algorithm selection process for a transfer learning environment in the absence of reliable validation techniques.

Multicue based object tracking frameworks have been extensively explored due to their numerous applications in the field of computer vision. However, the online adaptive fusion of multicue under scale and illumination variations, partial or full occlusion, background clutters and object deformation remains an open challenge problem. In order to address this, we propose an online visual tracking algorithm using adaptive integration of multicue in a particle filter framework. The particle level fusion process is modelled as Shafer’s model with a power set defined over two focal elements. Partial conflicting masses and conjunctive consensus among three cues are estimated for each evaluated particle. Partial conflicts among cues are redistributed using Dezert-Smarandache Theory (DSmT) based proportional conflict redistribution rules (PCR-6). Additionally, context sensitive transductive cues reliabilities are used for discounting the particle likelihoods for quick adaptation of tracker. In the proposed model, automatic boosting of good particles and suppression of low performing particles not only improves resampling process but also enhances tracker accuracy. Experimental validation over benchmarked video sequences reveals that the proposed multicue tracking framework outperforms state-of-the-art trackers under various dynamic environmental challenges.

Machine learning methods in software engineering are becoming increasingly important as they can improve quality and testing efficiency by constructing models to predict defects in software modules. The existing datasets for software defect prediction suffer from an imbalance of class distribution which makes the learning problem in such a task harder. In this paper, we propose a novel approach by integrating Over-Bagging, static and dynamic ensemble selection strategies. The proposed method utilizes most of ensemble learning approaches called Omni-Ensemble Learning (OEL). This approach exploits a new Over-Bagging method for class imbalance learning in which the effect of three different methods of assigning weight to training samples is investigated. The proposed method first specifies the best classifiers along with their combiner for all test samples through Genetic Algorithm as the static ensemble selection approach. Then, a subset of the selected classifiers is chosen for each test sample as the dynamic ensemble selection. Our experiments confirm that the proposed OEL can provide better overall performance (in terms of G-mean, balance, and AUC measures) comparing with other six related works and six multiple classifier systems over seven NASA datasets. We generally recommend OEL to improve the performance of software defect prediction and the similar problem based on these experimental results.

Class-imbalance is very common in real world. However, conventional advanced methods do not work well on imbalanced data due to imbalanced class distribution. This paper proposes a simple but effective Hybrid-based Ensemble (HE) to deal with two-class imbalanced problem. HE learns a hybrid ensemble using the following two stages: (1) learning several projection matrixes from the rebalanced data obtained by under-sampling the original training set and constructing new training sets by projecting the original training set to different spaces defined by the matrixes, and (2) undersampling several subsets from each new training set and training a model on each subset. Here, feature projection aims to improve the diversity between ensemble members and under-sampling technique is to improve generalization ability of individual members on minority class. Experimental results show that, compared with other state-of-the-art methods, HE shows significantly better performance on measures of AUC, G-mean, F-measure and recall.

The application of deep learning techniques in natural language processing tasks has been increased in recent years. Many studies have used the deep learning techniques to obtain a distributed representation of features. In particular, the convolutional neural network (CNN) with the distributed representation have subsequently been shown to be effective for the natural language processing tasks. This paper presents how to apply the CNN to speech-act classification. Then we analyze the experimental results on two issues, how to solve two problems about sparse speech-acts in train data and out of vocabulary, and how to utilize the advantages of CNN in the speech-act classification. As a result, we obtain the significant improved performances when CNN is applied to the speech-act classification.

This article has been retracted due to partial plagiarism and the author Murali Mohanan takes full responsibility for the partial plagiarism detected. A notice of retraction is published in Vol. 27 No. 8

This paper presents a novel differential evolution algorithm to solve dynamic optimization problems. In the proposed algorithm, the entire population is composed of several subpopulations, which are evolved independently and excluded each other by a predefined Euclidian-distance. In each subpopulation, the “DE/best/1” mutation operator is employed to generate a mutant individual in this paper. In order to fully exploit the newly generated individual, the selection operator was extended, in which the newly generated trial vector competed with the worst individual if this trial vector was worse than the target vector in terms of the fitness. Meanwhile, this trial vector was stored as the historical information, if it was better than the worst individual. When the environmental change was detected, some of the stored solutions were retrieved and expected to guide the reinitialized solutions to track the new location of the global optimum as soon as possible. The proposed algorithm was compared with several state-of-the-art dynamic evolutionary algorithms over the representative benchmark instances. The experimental results show that the proposed algorithm outperforms the competitors.

Many recommenders compute predictions by inferring the users’ preferences. However, in some cases, such as in e-education, the recommendations of pedagogical resources should rather be based on users’ memory. In order to estimate in real time and with low involvement what has been recalled by users, we designed a user study to highlight the link between gaze features and visual memory. Our protocol consisted in asking different subjects to remember a large set of images. During this memory test, we collected about 19 000 fixation points. Among other results, we show in this paper a strong correlation between the relative path angles and the memorized items. We then applied various classifiers and showed that it is possible to predict the users’ memory status by analyzing their gaze data. This is the first step so as to provide recommendations that fits users’ learning curve.