Computer Science - Conference Items
Permanent URI for this collection
Browse
Recent Submissions
Item Accessibility features of developmental platforms: Towards developing accessible mobile applications with cross-platform, research challenges and opportunities(DEIMS2024, 2024) Shoaib, Muhammad; Fitzpatrick, Donal; Pitt, Ian; Science Foundation IrelandMobile application development is a process of designing and developing smartphone-based applications for several platforms i.e., Android, iOS, etc. Mobile application development platforms can be categorised as native and cross-platform, each offering resources, frameworks, and tools to developers so they can develop smartphone-based applications using them. Native platforms concentrate on specific applications designed expressly for a particular platform i.e., Android or iOS. Crossplatform allows developers to develop applications with a single codebase that can work across multiple platforms. Examples include Xamarin and Unity. In this study, we have compared the accessibility features offered by Xamarin and Unity with those offered by native platforms. We analyzed the accessibility features offered by the iOS and Android platforms to determined whether and to what extent Xamarin and Unity offer the same accessibility capabilities. This analysis shows that numerous functionalities are shared by native iOS and Android APIs, however, some of them are not included in Xamarin and Unity. They also do not provide the implementation of fundamental APIs. This will require additional work by developers to write platformspecific code in native APIs to access the unavailable APIs. We have provided a comparative analysis of these platforms. Important accessibility aspects are highlighted that may prove useful for researchers and developers who are working to create accessible apps.Item Remote learning of mathematics for visually impaired students during COVID-19: Exploring online intervention, resources, challenges and issues(Springer Nature, 2024-01-30) Shoaib, Muhammad; Fitzpatrick, Donal; Pitt, Ian; Science Foundation IrelandDue to COVID-19, several dramatic changes have appeared all over the world i.e., travel restrictions, healthcare shortages, self-isolation, economic crises, social distancing, increases in food demand, job losses and closure of educational institutions. This led to a rapid transition from face-to-face to remote teaching. Most remote learning platforms are designed primarily for sighted students and are less useful for visually-impaired students. Especially in mathematics, it's quite difficult for visually-impaired students to access rich visual information, such as graphs, algebraic notations, geometric shapes and statistical formulas. This article provides an overview of mathematics learning resources during the COVID-19 pandemic. It explores assistive technology-based interventions which enable visually-impaired and blind students to actively participate in remote teaching and enhance their learning skills. It was noted that solutions are available for various platforms (i.e., smartphones, tablets, laptops & desktop PCs) to support visually-impaired students. Assistive Technology in remote learning also plays a key role in enhancing the mathematical skills of visually-impaired students. Furthermore, we have investigated the important COVID-19-related resources of different institutes which are very beneficial for the online education of visually-impaired students. Moreover, some challenges and issues are discovered such as internet connectivity, lack of interaction, inadequate support, one-way communication and less effective learning. Finally, some future directions are suggested for research i.e., parents should be a part of remote learning.Item RRI Adaptive: A standards compliant approach for equitable and stable congestion control in C-V2X networks(Institute of Electrical and Electronics Engineers (IEEE), 2023-12-22) McCarthy, Brian; O’Driscoll, Aisling; Science Foundation IrelandA key topic of interest in all vehicular networking technologies is their ability to deal with dense radio environments as widespread deployment becomes a reality. This requires congestion control mechanisms to maintain adequate communication performance. The current Cellular Vehicle-to-Everything (C-V2X) and its successor New Radio V2X (NR-V2X) standardised congestion control approach is table-based, utilising packet dropping. As shown in past works by the authors of this paper, these approaches exhibit congestion instability and require extensive configuration. To overcome this, algorithmic approaches were standardised, namely ETSI’s DCC Adaptive. Although this can be effective for wireless vehicular communications, it cannot be applied directly to NR-V2X/C-V2X due to incompatibility with the underlying radio scheduling approach, Sensing-Based Semi-Persistent Scheduling (SB-SPS). In previous work, the authors of this paper investigated in detail why this is the case and proposed an algorithmic approach that was compatible with the SB-SPS scheduler, namely RRI Adaptive . This paper provides an in-depth evaluation of RRI Adaptive . Importantly, its efficacy is evaluated not simply from the perspective of maintaining a desired channel load, but also from the perspective of maintaining effective application quality of service. This paper also describes the first study of fairness and stability in the context of C-V2X/NR-V2X congestion control. These are of increased importance, given dynamic channel conditions in vehicular scenarios, and reduced awareness due to degraded quality of service or vehicles starved of radio resources, may increase the likelihood of collisions.Item BHO-MA: Bayesian hyperparameter optimization with multi-objective acquisition(Springer, 2023-09) Dogan, Vedat; Prestwich, Steven; Science Foundation IrelandGood hyperparameter values are crucial for the performance of machine learning models. In particular, poorly chosen values can cause under- or overfitting in regression and classification. A common approach to hyperparameter tuning is grid search, but this is crude and computationally expensive, and the literature contains several more efficient automatic methods such as Bayesian optimization. In this work, we develop a Bayesian hyperparameter optimization technique with more robust performance, by combining several acquisition functions and applying a multi-objective approach. We evaluated our method using both classification and regression tasks. We selected four data sets from the literature and compared the performance with eight popular methods. The results show that the proposed method achieved better results than all others.Item An efficient non-Bayesian approach for interactive preference elicitation under noisy preference models(Springer, 2023-09-19) Pourkhajouei, Samira; Toffano, Federico; Viappiani, Paolo; Wilson, Nic; Science Foundation Ireland; European Regional Development Fund; Horizon 2020