Ayham Alomari, Ahmad Sami Al-Shamayleh, Norisma Idris, Aznul Qalid Md Sabri, Izzat Alsmadi, and Danah Omary
IEEE Access, Vol 11, Pp 112483-112501 (2023)
Abstractive summarization, novelty, warm-started models, deep learning, metrics, Electrical engineering. Electronics. Nuclear engineering, and TK1-9971
Abstractive summarization is distinguished by using novel phrases that are not found in the source text. However, most previous research ignores this feature in favour of enhancing syntactical similarity with the reference. To improve novelty aspects, we have used multiple warm-started models with varying encoder and decoder checkpoints and vocabulary. These models are then adapted to the paraphrasing task and the sampling decoding strategy to further boost the levels of novelty and quality. In addition, to avoid relying only on the syntactical similarity assessment, two additional abstractive summarization metrics are introduced: 1) NovScore: a new novelty metric that delivers a summary novelty score; and 2) NSSF: a new comprehensive metric that ensembles Novelty, Syntactic, Semantic, and Faithfulness features into a single score to simulate human assessment in providing a reliable evaluation. Finally, we compare our models to the state-of-the-art sequence-to-sequence models using the current and the proposed metrics. As a result, warm-starting, sampling, and paraphrasing improve novelty degrees by 2%, 5%, and 14%, respectively, while maintaining comparable scores on other metrics.
Rasheed Mohammad Nassr, Abdulaziz Aborujilah, Danah Ahmed Aldossary, and Alia Ahmed Abdullah Aldossary
Ieee Access IEEE Access
Medical education, Science - General, General Computer Science, Coronavirus disease 2019 (COVID-19), Process (engineering), business.industry, Online learning, online learning, General Engineering, Malaysia, Computers and Information Processing, Plan (drawing), 030218 nuclear medicine & medical imaging, Education, COVID-19 epidemic, Blended learning, 03 medical and health sciences, 0302 clinical medicine, General Materials Science, The Internet, 030212 general & internal medicine, business, Psychology, and country lockdown
The COVID-19 pandemic has struck the world and forced countries to go into lockdown including education sector. Students have been staying in hostels or houses, unable to go to university campuses. This situation has left university administrators no choice, but to have an online learning channel. Malaysian universities in particular have gone through many challenges to bring their online learning system up and ready to resume education process. However, students have found themselves caught in this situation (pure online learning) with no plan or readiness. Literature reviews showed that students encountered some challenges that could not be easily resolved. This study explored the challenges encountered by students of a government-linked university. This university is one of the largest in Malaysia with over 10 campuses across the country. This study collected 284 valid answers. The findings show that respondents lacked full readiness in this situation physically, environmentally, and psychologically with some differences in perspectives according to their gender, age, and residing state. Respondents were concerned about the implications of lockdown on their performance. The findings of this study indicate that a sudden switch to a pure online alternative creates considerable challenges to students who have no plans to be physically apart from classes. The findings also indicate that the current blended learning process which uses online learning as a support mechanism for face-to-face learning has faced a considerable challenge to replace it, particularly with unprepared students.
The goal of ecommerce recommender systems is to increase sales by providing accurate recommendations to the users. However, users purchase intention plays a vital role in making a purchase and ignoring their desire to make a purchase before providing a recommendation can negatively influence recommendation accuracy because it does not target the specific users. The proposed model identifies eight factors that can influence a user’s purchase intention and categorises them as user and population behaviors. These factors are fed into the AdaBoost algorithm to predict a user’s purchase intention. The results show that we achieve high accuracy and the ROC area is 0.91, when we use the identified factors to predict a user’s purchase intention. The proposed model outperforms two state-of-the-art models that have been used to predict a user’s purchase intention.
Amer Braik, Asaad Makhalfih, Tamer Kahtib, and Danah Barakat
2017 14th International Conference on Smart Cities: Improving Quality of Life Using ICT & IoT (HONET-ICT).
Optimal design, Electric power system, Engineering, business.industry, Photovoltaic system, business, Demand load, Sizing, Simulation, Power (physics), Block (data storage), Graphical user interface, and Reliability engineering
By the trend of utilizing photovoltaic (PV) systems, especially standalone PV system, in Palestine, the design of such system was carried out mostly using an intuitive method. Putting the intuitive method under the spot reveals a lot of its weaknesses regarding the over or under design results. However, many methods and algorithms were proposed to overcome the drawbacks of the intuitive method. Throughout this paper, we will introduce an algorithm to find the optimal design in terms of size of PV panels and battery block. The new method is mainly based on numerical basis and deals with hourly data of global radiation, demand load and ambient temperature. The proposed method has proved, as demonstrated next, its effectiveness to have results much better than the intuitive one. Such results enable the system to be either less cost system or more suitable and stable one in terms of deficit power. It has proved also the accurate prediction of the system's performance that is hourly performance of system attitude. To utilize the proposed method for stand-alone systems sizing, the method was fostered by Graphical User Interface. Which enables others to use it in order to get more effective results than intuitive method.
Danah Veronica Torres, Roel Ocampo, Jan Vincent Santos, John Norman Doctor, Monica Kathrina Villanoy, and Cedric Angelo M. Festin
Arc (geometry), Disaster risk reduction, SIMPLE (military communications protocol), Computer science, business.industry, Application framework, Mobile development, Mobile telephony, Routing (electronic design automation), business, and Computer network
In disaster situations communication and power infrastructure could be damaged, disrupting their associated services. This could delay relief efforts and cost more lives. In situations where connections are damaged and unreliable, peer-to-peer (P2P) and delay-tolerant networking (DTN) technologies can help provide alternative communication modes particularly in support of rescue and relief operations. Combining the benefits of the store-and-forward capability of DTN and the data propagation techniques of Wi-Fi Direct, we have created a flexible framework called ARC which allows users to create applications that could meet their specific requirements, particularly those in support of emergency communications. To demonstrate our framework, we created a simple application using ARC, and used this application to evaluate its DTN and Wi-Fi Direct group recovery capabilities.
2013 5th International Conference on Computer Science and Information Technology.
IT risk management, Risk management plan, IT risk, Actuarial science, Enterprise risk management, business.industry, Computer science, Risk analysis (business), Financial risk management, Risk factor (computing), business, and Risk management
Risk Management is one of the key cares of any organization strategic management. Referring to all the activities on-going in the organizations, risks are brought in for accomplishing the activities and their targets beneficially in all portfolios. Proper benefit of risk management is finding risks and their solutions. The result from a good risk management would be the maximization of the organization's value, it will also improve the chances of success than failure with risk assessing. Almost all risk reduction activities acquire cost of all probably applicable activities typically far overcomes the resources available. Hence there is the need to cleverly pick from among those activities to achieve at a cost-effective selection. In this paper we will suggest Strategic Risk Reduction (SRR) technique for producing optimal risk reduction strategies; we will reduce risk exposure for expected income by allowing several countermeasures per risk rather than one countermeasure per risk as previous work did.
Stefan E. Funk, Jedrzej Rybicki, Danah Tonne, and Peter Gietz
Service (systems architecture), Hypertext Transfer Protocol, Network security, business.industry, Computer science, computer.internet_protocol, Hypermedia, Virtualization, computer.software_genre, law.invention, World Wide Web, Upload, law, Server, Distributed data store, business, computer, and Humanities
Sustainable management of large amounts of research data is gaining in importance for research projects all over the world. The European project DARIAH aims to address this topic for the arts and humanities community. The DARIAH Bit Preservation, as a part of an archiving system for the arts and humanities, allows for a high performance, sustainable, and distributed storage of research data as basis of virtual research environments. A great challenge in designing such a service is to provide a standardized, consistent yet easy-to-use API for accessing the data that remains stable even if backend technology changes over time. As a solution, this paper presents the RESTful API of the DARIAH Bit Preservation which includes an administrative extension, and which is secured by an Authentication and Authorization Infrastructure (AAI) based on SAML. An exemplary implementation illustrates that the API offers distributed access by usage of the HTTP protocol and is able to handle a high number of files. Data transfer rates of up to 45 MB/s were achieved for uploading large files in the local network.