International Journal on Artificial Intelligence Tools

This study considers the Dynamic Minimum Cost Hybrid Berth Allocation Problem (DMCHBAP) with fixed handling times of vessels. The objective function to be minimized consists of three components: costs of positioning, waiting, and tardiness of completion for all vessels. A mathematical formulation of DMCHBAP, based on Mixed Integer Linear Programming (MILP), is proposed and used within the framework of commercial CPLEX 12.3 solver. As the speed of finding high-quality solutions is of crucial importance for an efficient and reliable decision support system in container terminal, two population-based metaheuristic approaches to DMCHBAP are proposed: combined Genetic Algorithm (cGA) and improvement-based Bee Colony Optimization (BCOi). Both cGA and BCOi are evaluated and compared against each other and against state-of-the-art solution methods for DMCHBAP on five sets of problem instances. The conducted computational experiments and statistical analysis indicate that population-based metaheuristic methods represent promising approaches for DMCHBAP and similar problems in maritime transportation.

Most pattern mining techniques almost singularly focus on identifying frequent patterns and very less attention has been paid to the generation of rare patterns. However, in several domains, recognizing less frequent but strongly related patterns have greater advantage over the former ones. Identification of compelling and meaningful rare associations among such patterns may proved to be significant for air quality management that has become an indispensable task in today’s world. The rare correlations between air pollutants and other parameters may aid in restricting the air pollution to a manageable level. To this end, efficient and competent rare pattern mining techniques are needed that can generate the complete set of rare patterns, further identifying significant rare association rules among them. Moreover, a notable issue with databases is their continuous update over time due to the addition of new records. The users requirement or behavior may change with the incremental update of databases that makes it difficult to determine a suitable support threshold for the extraction of interesting rare association rules. This paper, presents an efficient rare pattern mining technique to capture the complete set of rare patterns from a real environmental dataset. The proposed approach does not restart the entire mining process upon threshold update and generates the complete set of rare association rules in a single database scan. It can effectively perform incremental mining and also provides flexibility to the user to regulate the value of support threshold for generating the rare patterns. Significant rare association rules representing correlations between air pollutants and other environmental parameters are further extracted from the generated rare patterns to identify the substantial causes of air pollution. Performance analysis shows that the proposed method is more efficient than existing rare pattern mining approaches in providing significant directions to the domain experts for air pollution monitoring.

In the context of reinforcement learning we introduce the concept of criticality of a state, which indicates the extent to which the choice of action in that particular state influences the expected return. That is, a state in which the choice of action is more likely to influence the final outcome is considered as more critical than a state in which it is less likely to influence the final outcome.

We formulate a criticality-based varying step number algorithm (CVS) — a flexible step number algorithm that utilizes the criticality function provided by a human, or learned directly from the environment. We test it in three different domains including the Atari Pong environment, Road-Tree environment, and Shooter environment. We demonstrate that CVS is able to outperform popular learning algorithms such as Deep Q-Learning and Monte Carlo.

With the increase of cyber-attacks on millions of Internet of Things (IoT) devices, the poor network security measures on those devices are the main source of the problem. This article aims to study a number of these machine learning algorithms available for their effectiveness in detecting malware in consumer internet of things devices. In particular, the Support Vector Machines (SVM), Random Forest, and Deep Neural Network (DNN) algorithms are utilized for a benchmark with a set of test data and compared as tools in safeguarding the deployment for IoT security. Test results on a set of 4 IoT devices exhibited that all three tested algorithms presented here detect the network anomalies with high accuracy. However, the deep neural network provides the highest coefficient of determination R², and hence, it is identified as the most precise among the tested algorithms concerning the security of IoT devices based on the data sets we have undertaken.

Contextual multi-armed bandit algorithms are an effective approach for online sequential decision-making problems. However, there are limited tools available to support their adoption in the community. To fill this gap, we present an open-source Python library with context-free, parametric and non-parametric contextual multi-armed bandit algorithms. The MABWiser library is designed to be user-friendly and supports custom bandit algorithms for specific applications. Our design provides built-in parallelization to speed up training and testing for scalability with special attention given to ensuring the reproducibility of results. The API makes hybrid strategies possible that combine non-parametric policies with parametric ones, an area that is not explored in the literature. As a practical application, we demonstrate using the library in both batch and online simulations for context-free, parametric and non-parametric contextual policies with the well-known MovieLens data set. Finally, we quantify the performance benefits of built-in parallelization.

Inferring Gene Regulatory Networks from high-throughput gene expression data is a challenging problem, addressed by the systems biology community. Most approaches that aim at unraveling the gene regulation mechanisms in a data-driven way, analyze gene expression datasets to score potential regulatory links between transcription factors and target genes. So far, three major families of approaches have been proposed to score regulatory links. These methods rely respectively on correlation measures, mutual information metrics, and regression algorithms. In this paper we present a new family of data-driven inference methods. This new family, inspired by the regression-based paradigm, relies on the use of classification algorithms. This paper assesses and advocates for the use of this paradigm as a new promising approach to infer gene regulatory networks. Indeed, the development and assessment of five new inference methods based on well-known classification algorithms shows that the classification-based inference family exhibits good results when compared to well-established paradigms.

Most multi-label learning (MLL) techniques perform classification by analyzing only the physical features of the data, which means they are unable to consider high-level features, such as structural and topological ones. Consequently, they have trouble to detect the semantic meaning of the data (e.g., formation pattern). To handle this problem, a high-level framework has been recently proposed to the MLL task, in which the high-level features are extracted using the analysis of complex network measures. In this paper, we extend that work by evaluating different combinations of four complex networks measures, namely clustering coefficient, assortativity, average degree and average path length. Experiments conducted over seven real-world data sets showed that the low-level techniques often can have their predictive performance improved after being combined with high-level ones, and also demonstrated that there is no a unique measure that provides the best results, i.e., different problems may ask for different network properties in order to have their high-level patterns efficiently detected.