Artificial intelligence

Human Activity Recognition (HAR) plays a crucial role in various applications, including healthcare, fitness tracking, security, and smart environments, by enabling the automatic classification of human actions based on sensor and visual data. This dissertation presents a comprehensive exploration of HAR utilizing machine learning, sensor-based data, and Fusion approaches. HAR involves classifying human activities over time by analyzing data from sensors such as accelerometers and gyroscopes. Recent advancements in computational technology and sensor availability have driven significant progress in this field, enabling the integration of these sensors into smartphones and other devices. The first study outlines the foundational aspects of HAR and reviews existing literature, highlighting the importance of machine learning applications in healthcare, athletics, and personal use. In the second study, the focus shifts to addressing challenges in handling large-scale, variable, and noisy sensor data for HAR systems. The research applies machine learning algorithms to the KU-HAR dataset, revealing that the LightGBM classifier outperforms others in key performance metrics such as accuracy, precision, recall, and F1 score. This study underscores the continued relevance of optimizing machine learning techniques for improved HAR systems. The study highlights the potential for future research to explore more advanced fusion techniques to fully leverage different data modalities for HAR. The third study focuses on overcoming common challenges in HAR research, such as varying smartphone models and sensor configurations, by employing data fusion techniques.

Artificial intelligence is now a way of life meaning, it is hard to find any type of technology or technological advance that isn’t assisted by or powered by artificial intelligence and machine learning. From Siri on our iPhones to our computer tailored Netflix home screens to fast learning computerized and independent floor vacuums AI is everywhere you turn intruding on every aspect of daily functioning. As the pressure of said intrusion increases questions arise about whether all these advances can become crushing to humans. In some instances technology with AI components has been used to replace certain skill sets affecting the availability of employment surround jobs including, cashiers, hotel reception, customer service, taxi drivers, toll booths. And what about graphic design? Can a machine programmed with AI replace the creativity of a human spirit?
The research explores the tension between automated (artificial intelligence + machine learning) and manual, human initiated methods and practices in graphic design…
Can humans be removed from the process of graphic design? Expected outcome: No
How can the case study exploration coupled with the examination of certain considerations including ethical practices, human creativity, quality and originality demonstrate the necessity of human involvement.

This study applies the Response Escalation Paradigm (REP) to examine jealousy scores across five stages of increasing relationship threat (Huelsnitz et al., 2018). Participants rated their jealousy in scenarios involving a romantic partner with either a human or AI rival in virtual reality (VR), among other conditions. Consistent with hypothesis 1, jealousy rates increased across the stages in each condition. Consistent with hypothesis 2, people reported higher jealousy for the human rival in VR compared to the AI rival in VR at Stage 1. Inconsistent with hypothesis 3, participants did not experience faster escalation (the rate of increase in jealousy as the level of threat increases across the scenarios) for the human rival in VR relative to the AI rival in VR. Exploratory analyses did not reveal significant gender differences in jealousy responses. Overall, results indicate that individuals feel jealous about AI rivals as they do with human rivals.

The aim of this dissertation is to develop a comprehensive framework for designing optimal AI/ML-driven waveform solutions to achieve autonomous interference avoidance in fixed frequency bands. In the age of advanced wireless communications, minimizing interference is critical for maximizing the signal-to-interference-plus-noise ratio (SINR), particularly in densely occupied frequency environments. The research presented here focuses on developing adaptive MIMO waveform optimization techniques that dynamically adjust to varying interference conditions, enhancing communication reliability and performance for future autonomous machine-to-machine (M2M) networks. In addition to the established adaptive MIMO waveform optimization techniques, this dissertation investigates the implementation of AI-enhanced methods, to improve real-time adaptability in interference-rich environments. By leveraging neural networks, the framework enables the MIMO system to autonomously learn optimal waveform adjustments, providing resilience and efficiency under unpredictable interference conditions. This approach is validated through extensive simulations and experimental setups, demonstrating significant gains in SINR and overall communication reliability, marking a robust advancement toward achieving fully autonomous interference-avoiding communication in 6G and beyond networks. The AI-driven techniques further enhance the adaptability of the MIMO system across diverse interference scenarios, contributing to more consistent performance. These improvements offer a scalable approach for interference avoidance, adaptable to various network configurations.

The aim of this dissertation is to achieve a thorough understanding and develop an algorithmic framework for a crucial aspect of autonomous and artificial intelligence (AI) systems: Data Analysis. In the current era of AI and machine learning (ML), ”data” holds paramount importance. For effective learning tasks, it is essential to ensure that the training dataset is accurate and comprehensive. Additionally, during system operation, it is vital to identify and address faulty data to prevent potentially catastrophic system failures. Our research in data analysis focuses on creating new mathematical theories and algorithms for outlier-resistant matrix decomposition using L1-norm principal component analysis (PCA). L1-norm PCA has demonstrated robustness against irregular data points and will be pivotal for future AI learning and autonomous system operations.
This dissertation presents a comprehensive exploration of L1-norm techniques and their diverse applications. A summary of our contributions in this manuscript follows: Chapter 1 establishes the foundational mathematical notation and linear algebra concepts critical for the subsequent discussions, along with a review of the complexities of the current state-of-the-art in L1-norm matrix decomposition algorithms. In Chapter 2, we address the L1-norm error decomposition problem by introducing a novel method called ”Individual L1-norm-error Principal Component Computation by 3-layer Perceptron” (Perceptron L1 error). Extensive studies demonstrate the efficiency of this greedy L1-norm PC calculator.

The thesis focuses on the relationship between artificial intelligence (AI) and graphic design, aiming to understand AI's potential while emphasizing the unique value of human designers. This exploration will involve transforming AI-generated patterns into physical forms using paper, demonstrating the positive possibilities of AI in creative pursuits. By blending AI output with human creativity, the thesis will show how both play essential, distinct roles.
The study addresses concerns about AI's potential to replace human artists. By highlighting AI as a tool that enhances, rather than replaces, human creativity, the thesis debunks the misconception that AI will eradicate human artistry. Instead, AI can inspire and expedite creative processes while artists remain responsible for the final production.
Using Midjourney, a generative AI system, the research underscores AI's limitations in fully understanding or replicating human imagination. The study emphasizes the importance of human touch in design, particularly when using paper—a material with deep historical and cultural roots in graphic design and communication.
Furthermore, paper's use extends beyond art and design, proving its versatility and importance across various fields. The thesis highlights human ingenuity's role in maximizing paper's potential across disciplines, underscoring the critical contribution of human creativity.

At the site of neuronal communication, multiple interacting components drive synapse structure and function. Synaptic vesicle pools, membrane proteins, mitochondria, and perisynaptic astrocyte processes (PAPs) are all structures that can be altered through naturally occurring plasticity mechanisms to modulate neurotransmission, and disruption of these structures can result in synapse dysfunction and disease. Due to the minute size of the synapse, electron microscopy (EM) remains the gold standard for ultrastructural characterization; however, due to the complexity of EM datasets, extraction of information has become a bottleneck which places limits on the amount of data that can be collected and analyzed. A need exists for easy-to-use workflows that automate and enhance analysis throughput, to keep up with the streams of image data that are able to be produced. Here, I develop the use of AI algorithms, correlative microscopy techniques, and novel structural analysis methods to characterize postsynaptic mitochondria, PAPs, synaptic vesicles, and integral membrane proteins and their impact on synapse structure and function. I show that both postsynaptic mitochondria and PAPs in the visual cortex are positioned to support synapse structure and function; cleavage of a synaptic adhesion molecule affects synaptic vesicle accumulation in the amygdala; and presynaptic voltage gated calcium channels aggregate near active zone machinery in the brainstem. In addition, I highlight the use of virtual reality as a fast and intuitive tool for the identification and isolation of individual neurites in 3D EM. Thus, my work establishes novel technical approaches for EM and advances our understanding of neuronal communication through original research of several synaptic components.

The major objective of this dissertation was to create a framework which is used for medical image diagnosis. In this diagnosis, we brought classification and diagnosing of diseases through an Artificial Intelligence based framework, including COVID, Pneumonia, and Melanoma cancer through medical images. The algorithm ran on multiple datasets. A model was developed which detected the medical images through changing hyper-parameters.
The aim of this work was to apply the new transfer learning framework DenseNet-201 for the diagnosis of the diseases and compare the results with the other deep learning models. The novelty in the proposed work was modifying the Dense Net 201 Algorithm, changing hyper parameters (source weights, Batch Size, Epochs, Architecture (number of neurons in hidden layer), learning rate and optimizer) to quantify the results. The novelty also included the training of the model by quantifying weights and in order to get more accuracy. During the data selection process, the data were cleaned, removing all the outliers. Data augmentation was used for the novel architecture to overcome overfitting and hence not producing false absurd results the computational performance was also observed. The proposed model results were also compared with the existing deep learning models and the algorithm was also tested on multiple datasets.

Recent successes of Deep Learning-powered AI are largely due to the trio of: algorithms, GPU computing, and big data. Data could take the shape of hospital records, satellite images, or the text in this paragraph. Deep Learning algorithms typically need massive collections of data before they can make reliable predictions. This limitation inspired investigation into a class of techniques referred to as Data Augmentation. Data Augmentation was originally developed as a set of label-preserving transformations used in order to simulate large datasets from small ones. For example, imagine developing a classifier that categorizes images as either a “cat” or a “dog”. After initial collection and labeling, there may only be 500 of these images, which are not enough data points to train a Deep Learning model. By transforming these images with Data Augmentations such as rotations and brightness modifications, more labeled images are available for model training and classification! In addition to applications for learning from limited labeled data, Data Augmentation can also be used for generalization testing. For example, we can augment the test set to set the visual style of images to “winter” and see how that impacts the performance of a stop sign detector.
The dissertation begins with an overview of Deep Learning methods such as neural network architectures, gradient descent optimization, and generalization testing. Following an initial description of this technology, the dissertation explains overfitting. Overfitting is the crux of Deep Learning methods in which improvements to the training set do not lead to improvements on the testing set. To the rescue are Data Augmentation techniques, of which the Dissertation presents an overview of the augmentations used for both image and text data, as well as the promising potential of generative data augmentation with models such as ChatGPT. The dissertation then describes three major experimental works revolving around CIFAR-10 image classification, language modeling a novel dataset of Keras information, and patient survival classification from COVID-19 Electronic Health Records. The dissertation concludes with a reflection on the evolution of limitations of Deep Learning and directions for future work.

The presence of artificial intelligence (AI) has incrementally increased in our lives since its introduction in the 1950s and has exponentially increased in the last decade. In medicine, AI holds the promise of providing complete panoramic views of a patient’s medical history, improving medical decision making, avoiding errors such as misdiagnosis and unnecessary procedures, interpretating tests and making treatment recommendations. In this study, I examine the influence of AI on decision-making behaviors and the changes to the professional institution of medicine. This paper links theories of institutional change and professions to further our understanding of the processes of change in response to emergent technology. Recognizing that the autonomy of decision making is central to the model of professional work, this study (1) shows how changes in decision-making processes are a driver of change in the institution of professions and (2) highlights how this impacts the professional role identity of health care providers which has implications for how medicine is taught and how diagnoses are made.