Artificial intelligence's "deep learning" discipline has taken off, revolutionizing a variety of industries, from computer vision and natural language processing to healthcare and finance. Deep learning has shown extraordinary effectiveness in resolving complicated issues, and it has a wide range of potential applications, from autonomous vehicles to healthcare. The purpose of the survey to study deep learning's present condition, including recent advancements, difficulties, and constraints since the subject is currently fast growing. The basic ideas of deep learning, such as neural networks, activation functions, and optimization algorithms, are first introduced. We next explore numerous topologies, emphasizing their distinct properties and uses, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and generative adversarial networks (GANs). Further concepts, applications, and difficulties of deep learning are all covered in this survey paper's thorough review. This survey aid the academics, professionals, and individuals who want to learn more about deep learning and explore its applications to challenging situations in the real world.

This high-dimensional data is becoming increasingly common in various sectors, such as the social sciences, biology, medicine, and finance. It is defined by datasets that include many characteristics or dimensions. This study explores the idea of high-dimensional data, the difficulties it presents, and how it affects prediction results. Developing strategies to extract useful information from high-dimensional data is crucial due to its many issues, including the curse of dimensionality, increased processing complexity, and poor interpretability. This article presents research that employs feature selection techniques to solve problems related to high-dimensional data. The crucial process of feature selection is determining which features are most relevant and keeping them while eliminating those that are unnecessary or redundant. This method seeks to increase prediction accuracy, decrease overfitting, and improve model interpretability by lowering the dimensionality of the data. The present study examines three distinct feature selection methods, including Filter (SU & RELIEF), Wrapper (GENETIC & SFS), and Hybrid (combining filter & wrapper), to choose relevant features from high-dimensional data. We use a real-world high-dimensional customer dataset from UCI to illustrate how well our suggested feature selection methods work with the Naive Bayes machine learning algorithm. We demonstrate how feature selection strategies, both with and without feature selection, lead to improved prediction outcomes in high-dimensional data settings via a number of experiments and evaluations. The results show that using feature selection enhances the accuracy of the predictions. In contrast to filter and wrapper techniques, hybrid FS selects the best feature set from the three FS models. Researchers and practitioners working with high-dimensional data may make better decisions using these insights, eventually improving the prediction models’ quality and applicability.

The working of the Internet of Things (IoT) ecosystem indeed depends extensively on the mechanisms of real-time data collection, sharing, and automatic operation. Among these fundamentals, wireless sensor networks (WSNs) are important for maintaining a countenance with their many distributed Sensor Nodes (SNs), which can sense and transmit environmental data wirelessly. Because WSNs possess advantages for remote data collection, they are severely hampered by constraints imposed by the limited energy capacity of SNs; hence, energy-efficient routing is a pertinent challenge. Therefore, in the case of clustering and routing mechanisms, these two play important roles where clustering is performed to reduce energy consumption and prolong the lifetime of the network, while routing refers to the actual paths for transmission of data. Addressing the limitations witnessed in the conventional IoT-based routing of data, this proposal presents an FL-oriented framework that presents a new energy-efficient routing scheme. Such routing is facilitated by the ADDQL model, which creates smart high-speed routing across changing scenarios in WSNs. The proposed ADDQL-IRHO model has been compared to other existing state-of-the-art algorithms according to multiple performance metrics such as energy consumption, communication delay, temporal complexity, data sum rate, message overhead, and scalability, with extensive experimental evaluation reporting superior performance. This also substantiates the applicability and competitiveness of the framework in variable-serviced IoT-oriented WSNs for next-gen intelligent routing solutions.

In cloud computing, task scheduling has become more and more important for maximizing resource allocation, cutting down on energy use, and enhancing system performance. However, traditional algorithms often fall short in dynamic environments, leading to performance inefficiencies. For cloud computing energy efficiency, this research suggests a hybrid task scheduling method that combines Transformer-based fine-tuning and Cuckoo Search (HCS) optimization. The integration of bio-inspired optimization (HCS) and Modified BERT-Based Transformer allows for dynamic task reallocation, enhancing both energy consumption and resource utilization while maintaining high Service Level Agreement (SLA) compliance. Fine-tuning with Transformer models enables adjustments to improve scheduling efficiency and minimize energy wastage. The results show that the proposed method reduced energy consumption to 0.430 J (from 0.440 J) while maintaining 100% SLA compliance and task completion rate. Resource utilization improved to 88.2%, with energy efficiency of 0.0043 J/task. This hybrid approach offers significant advantages over traditional methods, such as BENBO + Bi-LSTM (0.687 J), PSO-PGA (0.617 J), and GAECS (0.537 J), achieving a reduced energy consumption of 0.440 J for 15 PM with 20 VMs, making it an ideal solution for scalable cloud computing environments.