International Journal of Computational Intelligence and Applications

Deep neural network (DNN) has many advantages. Autonomous driving has become a popular topic now. In this paper, an improved stack autoencoder based on the deep learning techniques is proposed to learn the driving characteristics of an autonomous car. These techniques realize the input data adjustment and solving diffusion gradient problem. A Raspberry Pi and a camera module are mounted on the top of the car. The camera module provides the images needed for training the DNN. There are two stages in the training. In the pre-training process, an improved autoencoder is trained by the unsupervised learning mechanism, and the characterization of the track is extracted. In the fine-tuning stage, the whole network is trained according to the labeled data, and then this model learns the driving characteristics better according to the samples. In the experimental stage, the car will predict the action of the car by the trained model in the autonomous mode. The experiment exhibits the effectiveness of the proposed model. Compared with the traditional neural network, the improved stack autoencoder has a better generalization ability and faster convergence speed.

In this paper, an existing rule-based intrusion detection system (IDS) is made more intelligent through the application of machine learning. Snort was chosen as it is an open source software and though it was performing well, it showed false positives (FPs). To find the best performing machine learning algorithms (MLAs) to use with Snort so as to improve its detection, we tested some algorithms on three available datasets. Support vector machine (SVM) was chosen along with fuzzy logic and decision tree based on their accuracy. Combined versions of algorithms through ensemble SVM along with other variants were tried on the generated traffic of normal and malicious packets at 10Gbps. Optimized versions of the SVM along with firefly and ant colony optimization (ACO) were also tried, and the accuracy improved remarkably. Thus, the application of combined and optimized MLAs to Snort at 10Gbps worked quite well.

Most learning algorithms require the practitioner to manually set the values of many hyperparameters before the learning process can begin. However, with modern algorithms, the evaluation of a given hyperparameter setting can take a considerable amount of time and the search space is often very high-dimensional. We suggest using a lower-dimensional representation of the original data to quickly identify promising areas in the hyperparameter space. This information can then be used to initialize the optimization algorithm for the original, higher-dimensional data. We compare this approach with the standard procedure of optimizing the hyperparameters only on the original input.

We perform experiments with various state-of-the-art hyperparameter optimization algorithms such as random search, the tree of parzen estimators (TPEs), sequential model-based algorithm configuration (SMAC), and a genetic algorithm (GA). Our experiments indicate that it is possible to speed up the optimization process by using lower-dimensional data representations at the beginning, while increasing the dimensionality of the input later in the optimization process. This is independent of the underlying optimization procedure, making the approach promising for many existing hyperparameter optimization algorithms.

Numerous fuzzy time series (FTS) forecasting methods have been proposed in scientific literature and has achieved growing attention in practice. Most of them are based on modeling fuzzy relationships(FRs) of the past data. In this paper, a new hybrid forecasting model based on the concept of time-variant FR group (TV-FRG), particle swarm optimization technique (PSO) and refinement technique in the defuzzization stage is presented to forecast the enrolments of the University and the stock market. PSO technique is utilized to adjust for obtaining the effective length of each interval in the universe of discourse for the forecasting model. Most of the existing forecasting models simply ignore the repeated FRs without any proper justification or accept the number of recurrence of the FRs without considering the appearance history of these FRs in the grouping FRs process. Therefore, the appearance history of the fuzzy sets on the right-hand side of the FRs is considered to establish the FR groups, called the TV-FRGs. Furthermore, the high-order TV-FRGs are used in order to obtain more accurate forecasting results in the defuzzication stage. To calculate these high-order TV-FRGs values, a refined defuzzication technique is developed, and incorporated in the proposed model. To verify the effectiveness of the proposed model, two numerical simulations are examined with the case of University enrollments and Taiwan futures exchange (TAIFEX). The experimental results show that the proposed model achieves good forecasting results compared to other existing forecasting models based on the high-order FTS. These promising results bring a significant meaning for the future work on the development of FTS and PSO algorithm in real-world forecasting applications.

System-on-chip (SOC) has become a mainstream design practice that integrates intellectual property cores on a single chip. The SOC test scheduling problem maximizes the simultaneous test of all cores by determining the order in which various cores are tested. The problem is tightly coupled with the test access mechanism (TAM) bandwidth and wrapper design. This paper presents a strength Pareto evolutionary algorithm for the SOC test scheduling problem with the objective of minimizing the power-constrained test application time, wrapper design and TAM assignment in flat and hierarchical core-based systems. We demonstrate the effectiveness of the method using the ITC’02 benchmarks.

Semi-supervised boosting strategy aims at improving the performance of a given classifier with a multitude of unlabeled data. In semi-supervised boosting strategy, a similarity is needed to select unlabeled samples and then a pseudo label will be assigned to the unlabeled sample. A good similarity is helpful to assign a more proper pseudo label to unlabeled samples. Those selected samples with their pseudo labels will serve as labeled samples to train the new component classifier. So, similarity is important in semi-supervised boosting. Gaussian kernel similarity $S_{ij}=exp(-\frac{\parallel x_i-x_j{\parallel }^2}{2{\sigma }^2})$ is used in semi-supervised boosting strategy. There are two drawbacks, first, the Euclidean distance $\parallel x_i-x_j{\parallel }^2$ cannot characterize the complicated relationship between the data samples; second, the parameter $\sigma$ needs to set carefully. So, this paper proposes a novel adaptive similarity based on sparse representation for semi-supervised boosting. Our sparse representation is learned from a “clean” dictionary, which is a low rank matrix obtained from the sample matrix. We evaluate the proposed method on COIL20 databases. Experimental results show that: the semi-supervised boosting algorithm with sparse representation similarity outperforms the algorithm with Gaussian kernel similarity.