Telus (Canada)

Artificial intelligence is rapidly transitioning from a set of discrete analytical tools into a foundational enterprise infrastructure. The concept of AI factories captures this shift by framing GPU-powered systems as integrated environments where data ingestion, model training, inference, and continuous optimization occur at industrial scale. This abstract examines AI factories as enterprise intelligence infrastructure and analyzes how high-performance computing architectures are redefining value creation across multiple industries. Unlike traditional IT systems optimized for transactional efficiency, AI factories are designed for learning velocity, enabling organizations to transform raw data into predictive, adaptive, and autonomous capabilities. Central to this transformation are graphics processing units, high-bandwidth interconnects, and optimized software stacks that jointly support parallel computation, real-time analytics, and large-scale model deployment. The study synthesizes recent industry practices and academic perspectives to demonstrate how AI factories are reshaping sectors such as manufacturing, finance, healthcare, energy, and logistics. In manufacturing, GPU-accelerated intelligence enables predictive maintenance, process optimization, and digital twins that reduce downtime and resource waste. Financial institutions leverage AI factories for fraud detection, algorithmic trading, and personalized risk assessment at unprecedented speed and accuracy. In healthcare, these infrastructures support medical imaging analysis, clinical decision support, and population-level health forecasting, improving outcomes while lowering operational costs. Across energy and supply chains, AI factories facilitate demand forecasting, grid optimization, and resilient logistics planning under uncertainty. Beyond operational gains, the paper argues that AI factories represent a strategic asset that redefines organizational value creation. By embedding continuous learning into core operations, enterprises shift from reactive decision-making toward anticipatory and adaptive intelligence. This transition alters competitive dynamics, workforce skill requirements, and governance models, raising critical questions around data stewardship, ethical AI deployment, cybersecurity, and sustainability of energy-intensive computing systems. The abstract concludes by positioning AI factories as the backbone of next-generation enterprises, emphasizing that their successful adoption depends not only on GPU capacity, but on integrated strategy, responsible governance, and alignment with long-term business and societal objectives. Collectively, these insights position AI factories as scalable intelligence platforms capable of converting computational power into sustained innovation, measurable productivity, and resilient competitive advantage across complex digital ecosystems globally and.

This paper examines including Indigenous perspectives into data ethics strategies and artificial intelligence practices as part of a corporate reconciliation commitment. Through three workshops with First Nations, Inuit, and Métis Peoples, the research explored Indigenous Knowledge, data ethics, and responsible AI use.The study identified three key themes: respecting Indigenous data sovereignty, ensuring Indigenous involvement in AI development, and implementing distinctions-based approaches. Case studies demonstrate practical applications, including the OCAP® principles learning circle, AI image declaration, and partnerships with Indigenous-owned technology companies.Key findings emphasize the importance of distinctions-based approaches, including Indigenous data sovereignty principles, and involving Indigenous Peoples across multiple levels of AI development and governance. The findings conclude that prioritizing Indigenous involvement in emerging technologies is crucial for ethical innovation and advancing reconciliation while fostering sustainable, community-driven outcomes.

Accurate data is essential in modern healthcare, necessitating innovative strategies for data integrity. Patient health depends on reliable clinical data from healthcare entities. Current approaches prioritize confidentiality and availability, neglecting data integrity. Enhancing patient trust in clinical data integrity can improve evaluations of data-generating processes. Medical malpractice lawsuits can harm patient-physician relationships and lead to inconsistent records, causing data disputes. While recording interactions may help, US law restricts recordings of malpractice claims due to confidentiality. Such recordings must be sealed for twenty years and are seldom reviewed, resulting in inaccurate estimates. This paper advocates for the integrity of digitally recorded medical interactions for medico-legal purposes and assesses data integrity mechanisms’ health efficacy. It reviews managing recorded medical data and proposes three methods for real-time hospital recording and auditing. Objectives include developing genomics-based precision medicine, nationwide protocol detection, care assessment, litigation analysis, and privacy consent audits, although innovations have largely failed to yield actionable results.

Facilities for finishing cattle in North America include open lots with either earthen or solid pen floors, and confinement barns with either a bed pack or a deep pit. Confinement barn facili­ties are typically found in areas that receive a greater amount of precipitation, and actually represent a relatively small propor­tion of the total cattle on feed in North America. Currently, the most common feedlot facility type would be an open lot with an earthen floor. However, the use of roller compacted con­crete has been identified as an economically viable product for amending feedlot pen floors to create a solid surface on which cattle are housed. The objective is to identify the pros and cons of roller compacted concrete as a pen floor surface, and discuss the management and production implication of this pen floor surface technology so that feedlot practitioners are equipped to have an informed discussion with their clients about the incor­poration of this product into their production systems.

undergo post-mortem examination (Swedish Veterinary Agency, 2022).This means that the reported cause of death is often based on the farmers own assessment, which can be misleading or even incorrect (McConnel et al., 2009;Thomsen et al., 2012;Alvåsen et al., 2014).The limited numbers of post-mortem examinations of Swedish cattle might be due to logistical barriers, e.g., transport of the cadaver to a post-mortem site within a reasonable time.In Sweden, the cost of a post-mortem examination at a

Automatic human fall detection is a challenging task in healthcare. Using deep learning models in conjunction with cameras have been proven to be more accurate and efficient in detecting falls in comparison to wearable technology and other more intrusive alternatives. In this paper, we propose a long-term recurrent convolutional network (LRCN) architecture for human fall detection, which consists of a custom convolutional neural network specifically designed to address the challenges of this task and the limited amount of available data, followed by a long short-term memory (LSTM) neural network which decodes sequential information to determine if there has been a fall. Performance evaluations show that our LRCN architecture outperforms the existing state-of-the-art method by 11.20%.

One of the main advantages of Light Field (LF) technology is that it provides a truly immersive experience, critical for computer vision, autonomous driving and medical applications. However, one of the main problems of light field is the size of the data captured, which significantly increases bandwidth requirements. In this paper, we introduce a learning-based LF view synthesis approach for efficient transmission and storage, fundamental for performing remote surgery and storing data. This is achieved by dropping specific views at the transmitting end and then efficiently synthesizing them at the receiver end. Our deep learning approach uses the epipolar image plane (EPI) information to ensure smooth disparity between the generated and original views. We consider plenoptic, synthetic LF content and camera array implementations which support different baseline settings. Experimental results show that our proposed method outperforms state-of-the-art light field view synthesis techniques, offering improved visual quality for the generated views.

The COVID-19 pandemic has had a significant impact on patient lifestyles with new measures designed to reduce transmission of the virus drastically transforming life as we know it. Public health interventions — alongside advances in medical treatments, vaccine technology, and gene sequencing — have drastically reduced the impact of the pandemic across the world. Yet there has been relatively little focus on the potential role of physical activity (PA) in reducing disease burden during the pandemic. We discuss the latest evidence related to the role of exercise or physical pre-rehabilitation before infection and consider whether this may be an overlooked public health strategy. ### Living with COVID-19: ’Exercise is medicine.’ As we progress through the second year of the pandemic, there has been a renewed focus on adjustments that individuals and society will have to make as we continually adapt to life with COVID-19. Key public health messages regarding the importance of ’ hands, face, space and fresh air ‘ appear prominently on all government briefings but exercise, once famously quoted as being the ‘ miracle cure ’ by the Academy of Medical Royal Colleges, is conspicuously absent.1 Despite the evidence supporting the physical and mental health benefits of exercise at a population level, there remains a …