Subset selection is always a hot topic in the community of evolutionary multi-objective optimization (EMO) since it is used in mating selection, environmental selection, and final selection. In the first two scenarios...
详细信息
Software effort estimation is a vital component of project management, encompassing the prediction of time, cost, and resources necessary for software development. Accurate effort estimation plays a pivotal role in ef...
详细信息
Improving the generative and representational capabilities of auto-encoders is a hot research topic. However, it is a challenge to jointly and simultaneously optimize the bidirectional mapping between the encoder and ...
详细信息
Improving the generative and representational capabilities of auto-encoders is a hot research topic. However, it is a challenge to jointly and simultaneously optimize the bidirectional mapping between the encoder and the decoder/generator while ensuing convergence. Most existing auto-encoders cannot automatically trade off bidirectional mapping. In this work, we propose Bi-GAE, an unsupervised bidirectional generative auto-encoder based on bidirectional generative adversarial network (BiGAN). First, we introduce two terms that enhance information expansion in decoding to follow human visual models and to improve semantic-relevant feature representation capability in encoding. Furthermore, we embed a generative adversarial network (GAN) to improve representation while ensuring convergence. The experimental results show that Bi-GAE achieves competitive results in both generation and representation with stable convergence. Compared with its counterparts, the representational power of Bi-GAE improves the classification accuracy of high-resolution images by about 8.09%. In addition, Bi-GAE increases structural similarity index measure (SSIM) by 0.045, and decreases Fréchet inception distance (FID) by in the reconstruction of 512*512 images.
Phishing websites have become a significant cybersecurity threat, hosting malware and exploiting users by mimicking popular sites. Victims suffer financial loss, compromised privacy, and damaged reputation. Urgent sol...
详细信息
This research work presents a Empathic chatbot leveraging advanced natural language processing techniques. The chatbot employs a deep learning architecture, specifically a variant of the Transformer model, to generate...
详细信息
With the advent of transfer learning approaches, Natural Language Processing (NLP) problems have experienced tremendous progress, as demonstrated by models such as Generative Pre-trained Transformers (GPT) and Bidirec...
详细信息
ISBN:
(纸本)9798350388800
With the advent of transfer learning approaches, Natural Language Processing (NLP) problems have experienced tremendous progress, as demonstrated by models such as Generative Pre-trained Transformers (GPT) and Bidirectional Encoder Representations from Transformers (BERT). The usefulness of such transfer learning strategies across a range of NLP tasks and domains is investigated in this work. The study uses a methodical methodology to assess BERT and GPT's performance on a wide range of tasks. In addition, the study evaluates the generalizability and flexibility of these models across a broad variety of disciplines, including social media, finance, legal, and biological literature. The study's methodology entails rigorous assessment utilizing task-specific standard metrics after pre-trained BERT and GPT models have been fine-tuned using task-specific datasets. To determine the relative benefits and drawbacks of transfer learning strategies in various contexts, comparative studies are carried out against baseline models and other cutting-edge methodologies. Additionally, the study looks at how the performance of BERT and GPT is affected by variables including task difficulty, dataset size, and domain specificity. The results provide a comprehensive understanding of the benefits and drawbacks of transfer learning strategies in a variety of NLP tasks and domains. While BERT performs admirably on tests requiring semantic comprehension and contextual knowledge, GPT is superior at producing text that is both cohesive and appropriate to the situation. Both models, however, show sensitivity to dataset features and idiosyncrasies unique to the domain, indicating the necessity for customized fine-tuning techniques for best results. All things considered, this study advances our knowledge of the usefulness and efficiency of transfer learning strategies and provides insightful information for academics and practitioners who want to use BERT, GPT, and related models in a variety
In Vehicular Edge Computing (VEC), efficient task offloading is essential for reducing latency and enhancing system performance, especially in resource-constrained environments. In this regard, Roadside Units (RSUs) c...
详细信息
Dengue fever is a common vector-borne sickness in tropical regions, particularly in India, Bangladesh, and Pakistan. This disease, caused by mosquitoes, affects people of all ages in more than a hundred nations throug...
详细信息
Lung image compression and segmentation are essential diagnostic and monitoring techniques for a variety of respiratory ailments. We present a novel deep learning-based technique for segmenting and compressing lung im...
详细信息
Several digital dangers were investigated. Malware dominated analysis with 45 attacks. We found 30 phishing attacks. 22 data breaches, 15 cyber espionage, 18 identity theft. This indicates the kind and frequency of ha...
详细信息
暂无评论