Narrow Results

As a core component of the Space Telemetry, Track and Command (TT&C) system, the TT&C software’s quality is the key factor to ensure the successful implementation of space TT&C missions. Due to the complexity of space TT&C missions, incremental development is adopted in which frequent testing is required. Many problems of the original manual testing are exposed, such as large consumption of human resources, low efficiency of testing and quality highly relying on human expertise. Automatic testing method is urgently required. However, the testing of the TT&C software highly depends on the domain knowledge of space TT&C missions, which is complex and professional. This hinders the direct application of existing software automatic testing methods to the TT&C software. Therefore, in this paper, we propose an automatic testing method of the TT&C software based on domain knowledge. The domain knowledge description of the space TT&C domain is defined and a set of guidance principles for eliciting domain knowledge elicitation is given. Testing cases are automatically generated and executed by combining the domain knowledge and the image recognition results. Evaluations show that our method can realize the automatic testing of space TT&C software increments with higher accuracy. Its time cost is reduced by more than 50% compared with the manual testing, and will not increase rapidly with growing software maintenance scale. The time cost in domain knowledge elicitation will not affect the testing efficiency.

It is commonly understood that machine learning algorithms discover and extract knowledge based on data at hand. However, a huge amount of knowledge is available which is in machine-readable format and ready for inclusion in machine learning algorithms and models. In this paper, we propose a framework that integrates domain knowledge in form of ontologies/hierarchies into logistic regression using stacked generalization. Namely, relations from ontology/hierarchy are used in stacking manner in order to obtain higher, more abstract concepts. Obtained concepts are further used for prediction. The problem we solved is unplanned 30-days hospital readmission, which is considered as one of the major problems in healthcare. Proposed framework yields better results compared to Ridge, Lasso, and Tree Lasso Logistic Regression. Results suggest that the proposed framework improves AUC by up to 9.5% on pediatric datasets and up to 4% on morbidly obese patients’ datasets and also improves AUPRC by up to 5.7% on pediatric datasets and up to 2.6% on morbidly obese patients’ datasets on average. This indicates that the inclusion of domain knowledge improves the predictive performance of Logistic Regression.

As Artificial Intelligence (AI) systems become increasingly embedded in our daily lives, it is of utmost importance to ensure that they are both fair and reliable. Regrettably, this is not always the case for predictive policing systems, as evidence shows biases based on age, race, and sex, leading to wrongful identifications of individuals as potential criminals. Given the existing criticism of the system’s unjust treatment of minority groups, it becomes essential to address and mitigate this concerning trend. This study delved into the infusion of domain knowledge in the predictive policing system, aiming to minimize prevailing fairness issues. The experimental results indicate a considerable increase in fairness across all metrics for all protected classes, thus fostering greater trust in the predictive policing system by reducing the unfair treatment of individuals.

In this paper, we present a novel model for improving the performance of Domain Dictionary-based text categorization. The proposed model is named as Self-Partition Model (SPM). SPM can group the candidate words into the predefined clusters, which are generated according to the structure of Domain Dictionary. Using these learned clusters as features, we proposed a novel text representation. The experimental results show that the proposed text representation-based text categorization system performs better than the Domain Dictionary-based text categorization system. It also performs better than the system based on Bag-of-Words when the number of features is small and the training corpus size is small.

In this paper, we propose domain knowledge-based link prediction algorithm in customer-product bipartite network to improve effectiveness of product recommendation in retail. The domain knowledge is classified into product domain knowledge and time context knowledge, which play an important part in link prediction. We take both of them into consideration in recommendation and form a unified domain knowledge-based link prediction framework. We capture product semantic similarity by ontology-based analysis and time attenuation factor from time context knowledge, then incorporate them into network topological similarity to form a new linkage measure. To evaluate the algorithm, we use a real retail transaction dataset from Food Mart. Experimental results demonstrate that the usage of domain knowledge in link prediction achieved significantly better performance.

In this chapter, we propose to combine domain knowledge and deep learning in order to improve human activity recognition (HAR) models. We explore empirically two fundamental questions: (1) which knowledge to incorporate into HAR models and (2) how to incorporate it? We focus on the knowledge about the topological structure of on-body sensor-deployments and the mutual interactions that emerge among these sensors. We report on experiments conducted on the Sussex Huawei locomotion-transportation (SHL) dataset which features a structured data generation process. Obtained results open up perspectives for the development of deep learning-based HAR models which leverage domain knowledge in order to improve both robustness and data-efflciency.

Accurate prediction of highway traffic flow during holidays is crucial for traffic management and control. Traditional traffic flow prediction methods mainly rely on historical data and statistical models, which are difficult to cope with the complex and changeable traffic conditions during holidays. This chapter proposes a short-term highway traffic flow prediction method based on abductive learning for holidays. The method combines machine learning and logical reasoning, learns traffic flow features from historical data, and utilizes expert knowledge to construct logical rules to constrain and optimize the prediction results. Meanwhile, semi-supervised learning is introduced to improve the generalization ability of the model by using unlabeled data. Experimental results show that the method can effectively improve the accuracy of short-term high-way traffic flow prediction during holidays and provide reliable decision support for traffic management.