Services Computing – SCC 2019

Sensitive data is usually encrypted to protect against data leakage and unauthorized access for cloud storage services. Generally, the remote user has no knowledge of the actual data format stored in the cloud, even though a cloud server promises to store the data with encryption. Although a few works utilize data encapsulation and remote data checking to detect whether the sensitive data is protected securely in the cloud, they still suffer from a number of limitations, such as heavy computational cost at the user side and poor practicality, that would hinder their adoptions. In this paper, we propose a practical verification scheme to allow users to remotely evaluate the actually deployed data encryption protection in the cloud. We employ the pseudo-random number generator and present a data encapsulation solution, which can benefit users with significant cost savings. By imposing monetary rewards or penalties, our proposed scheme can help ensure that the cloud server stores data encrypted at rest honestly. Extensive experiments are conducted to further demonstrate the efficiency and practicality of the proposed scheme.

As the first level of a BPM strategy, being able to design event-oriented models of processes is a must-have competence for every modern business. Unfortunately, industrial procedures have reached a certain complexity making the designing task complex enough to discourage businesses facing the blank page. Moreover, the 21^st century witnesses the emergence of myriads of norms and external regulations that businesses want to abide by. Although domain experts have a limited process modelling and norm interpretation knowledge, they know how to describe their activities and their sequencing. With progresses made in the artificial intelligence, particularly in the natural language processing domain, it becomes possible to automatize the task of creating a process in compliance with norms. This paper presents a business-oriented prototype assisting users in getting certifiable specific business processes. We detail the metamodel used to separately model norms and business’ existing procedures and then, the algorithm envisaged to deduce a corresponding cartography of processes.

Quality of service (QoS) is a critical nonfunctional property and a criterion for the selection of web services (WSs); due to its importance, many QoS-aware or QoS-based approaches have been proposed and developed. However, with the existence of numerous approach-based studies of QoS of WSs, we consider that the deficiency in the existing research is the lack of a systematic investigation and analysis of real-world QoS data to discover and understand the characteristics of such data. Therefore, in this paper, we first define a number of research questions related to the properties of WSs’ QoS that could be interesting to WS/QoS researchers. Then, two real-world, large-scale QoS datasets are chosen, and a number of experiments that address the defined research questions are designed and performed on those datasets. Finally, based on the experimental results, the answer to each research question is discussed in detail.

The main contribution of this paper is to empirically reveal and confirm several useful and interesting properties of real-world QoS. For example, it is found that the distance between a service consumer and its invoked WS does not influence the invocation failure rates of the WSs; however, this distance is indeed correlated to the consumer-perceived WS performance in that a shorter distance can lead to a shorter response time and higher throughput (i.e., a better performance) of WSs according to our experimental results.

With the development of data mining and machine learning techniques, data-driven based technology-enhanced learning (TEL) has drawn wider attention. Researchers aim to use established or novel computational methods to solve educational problems in the ‘big data’ era. However, the readiness of data appears to be the bottleneck of the TEL development and very little research focuses on investigating the data scarcity and inappropriateness in the TEL research. This paper is investigating an emerging research topic in the TEL domain, namely micro learning. Micro learning consists of various technical themes that have been widely studied in the TEL research field. In this paper, we firstly propose a micro learning system, which includes recommendation, segmentation, annotation, and several learning-related prediction and analysis modules. For each module of the system, this paper reviews representative literature and discusses the data sources used in these studies to pinpoint their current problems and shortcomings, which might be debacles for more effective research outcomes. Accordingly, the data requirements and challenges for learning analytics in micro learning are also investigated. From a research contribution perspective, this paper serves as a basis to depict and understand the current status of the readiness of data sources for the research of micro learning.

In order to personalize users’ recommendations, it is essential to consider their personalized preferences on non-functional attributes during service recommendation. However, to increase recommendation accuracy, it is essential that recommendation systems include users’ evolving preferences. It is not sufficient to only consider users’ preferences at a point in time. Existing time-based recommendation systems either disregard rich and useful historical user invocation information, or rely on information from similar users, and thus, fail to thoroughly capture users’ dynamic preferences for personalized recommendation. This work proposes a method to personalize users’ recommendations based on their dynamic preferences on non-functional attributes. First, we compose a user’s preference profile as a time series of his/her invocation preference and pre-invocation dependencies (i.e. the different services that were viewed prior to invoking the preferred service). We model a user’s invocation preference as a combination of non-functional attribute values observed during service invocation, and topic distribution from WSDL of the invoked service using Hierarchical Dirichlet Process (HDP). Next, we employ long short-term memory recurrent neural networks (LSTM-RNN) to predict the user’s future invocation preference. Finally, based on the predicted future invocation preference, we recommend service(s) to that user. To evaluate our proposed method, we perform experiments using real-world service dataset, WS-Dream.

In this paper, we consider managing service performance starting from the composition time, aiming to reduce the risk of execution failures during service composition. We use ARIMA to predict workloads of the services at the time when they are likely to be invoked and subsequently predict the response time and chances that the requests for accessing the services may be declined due to admission control. The in-depth analysis can help avoid timing failures during service execution. However, these analyses may incur overhead and we introduce a two-phase composition algorithm to reduce the potential overhead. Our system also considers continuous monitoring and service recomposition to greatly increase the probability of completing the service execution within the deadline. Experimental results show that our service management approach can greatly improve the success rate for meeting the deadline.

The rise of chatbots in the finance sector is the latest disruptive force that has change the way customers interact. The adoption of Artificial Intelligence powered chatbots particularly in the banking industry has changed the face of communication interface between bank and customers. This paper explores the effectiveness of the current use of chatbot in Singapore’s banking industry. The banking sector in Singapore play a significant role in Singapore economy. It also investigates the current chatbot functionality to determine if it can meet the ever-changing expectation of customers.

We present a method of inspecting surface defects of ring magnets by using deep learning technology, and the inspection system developed utilizing this method has achieved much better accuracy and speed than human inspectors in actual production environment, while such accuracy and speed are essential for such systems. The proposed method can also be used for the surface defect inspection of many other industrial products and systems.