Search

Deep learning approaches are generally complex, requiring extensive computational resources and having high time complexity. Transfer learning is a state-of-the-art approach to reducing the requirements of high computational resources by using pre-trained models without compromising accuracy and performance. In conventional studies, pre-trained models are trained on datasets from different but similar domains with many domain-specific features. The computational requirements of transfer learning are directly dependent on the number of features that include the domain-specific and the generic features. This article investigates the prospects of reducing the computational requirements of the transfer learning models by discarding domain-specific features from a pre-trained model. The approach is applied to breast cancer detection using the dataset curated breast imaging subset of the digital database for screening mammography and various performance metrics such as precision, accuracy, recall, F1-score, and computational requirements. It is seen that discarding the domain-specific features to a specific limit provides significant performance improvements as well as minimizes the computational requirements in terms of training time (reduced by approx. 12%), processor utilization (reduced approx. 25%), and memory usage (reduced approx. 22%). The proposed transfer learning strategy increases accuracy (approx. 7%) and offloads computational complexity expeditiously., Competing Interests: The authors declare there are no competing interests., (©2024 Safdar Ali Khan et al.)

Despite the important role of the genre awareness in facilitating the effective communication in the global business context and the need to teach and practice writing for various social purposes, there was scant classroom-based research on effects of providing peer feedback on EFL learners' genre awareness in business writing for fulfilling the particular communicative purpose. This study examined whether student reviewers could improve their genre awareness in business letter writing by providing peer feedback to other students' writings at varying levels. Sixty business English majors, taking the business writing course for 1 year under the tutelage of the same instructor, participated in the current study. They were randomly assigned to the experimental group who reviewed the students' drafts at different levels (high, medium and low) and gave written comments and the control group who received no treatment but did self-revision. Both groups followed the genre-specific evaluation criteria during peer feedback and self-revision. Repeated-measures ANOVA on the two groups' writing performances at different timepoints (pretest, immediate posttest, transfer posttest and delayed transfer posttest) demonstrated that the participants in the experimental group had better performance than those in the control group at the transfer posttest as well as the delayed transfer posttest. Moreover, providing weakness comments on both the low-quality and the medium-quality drafts had significant effects on their own writing quality at the subsequent tests, whereas providing strength feedback on the high-quality drafts had statistically significant impacts on their own writing quality at the immediate posttest and the delayed transfer posttest. However, multiple linear regression analyses demonstrated that only offering weakness comments to the medium-quality drafts could effectively predict the reviewers' overall writing quality in the immediate posttest, the transfer posttest, and the delayed transfer posttest. Tentative research and pedagogical implications of the findings were discussed., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Wu, Zhu and Yang.)