Artificial Intelligence (AI) can provide humans great relief from numerous repetitive tasks, with automation increasing productivity. Furthermore, AI powered machines and devices are fast and efficient: learning, predicting, and deciding with superhuman accuracy.\nConcurrently, the use cases for augmentative AI are expanding, with numerous industries and organizations looking to tap into the potential. This is clearly the situation in two key Healthcare and Life Sciences (HCLS) areas: neuroscience and radiology.\nDeep learning and image recognition\nIn both areas, radiologists, MRI technicians and physicians must spend hours sifting through images, searching for markers and anomalies that may spur a disease diagnosis. Whether it be lung cancer or Alzheimer\u2019s, it could traditionally take hours or even days for a clear diagnosis. But now, with the advent of AI and the capabilities brought about by deep learning (DL) for image recognition, we\u2019re seeing clear advances and positive patient outcomes.\nLet\u2019s look at two recent research-based case studies involving AI augmentation and image recognition. These are both exceptional examples of human-machine interaction and collaboration, showcasing the possibilities of our continued co-existence and development.\nMcGill University and Courtois NeuroMod Project\nResearchers and technologists from McGill University and University of Montreal have come together in the Courtois NeuroMod project, aimed at building more robust AI systems that will eventually replicate human brain patterns.\nThe research team is collecting and analyzing massive datasets from a small group of volunteers who watch videos, look at images and play video games, all while inside an MRI machine. This allows researchers to track and record brain activity of these subjects. They then develop benchmarks that are essentially decoding the brain, paving a path to mimicking it in more versatile AI models and neural networks. In this way, they hope AI will behave more like humans in the transfer of knowledge.\nTo accomplish these tasks, the researchers needed leading-edge AI tools and advanced IT resources to accelerate brain mapping. They turned to Intel and Dell Technologies to provide the compute and storage infrastructure along with the data science and supercomputing resources of Dell Technologies HPC & AI Innovation Lab in Austin, Texas. And the Courtois NeuroMod team keeps driving groundbreaking brain research, opening new areas for AI, enabling processing of massive datasets, and accelerating the training of complex AI models.\nSURFsara\nRadiology is an exacting and time-consuming discipline, one that requires hours of assiduous analysis of X-ray images. It\u2019s clearly an area where AI can assist in lightening the workload. To this effort, scientists and researchers in the Netherlands turn to SURFsara, a national HPC center in the Netherlands that enhances its capabilities by collaborating with other organizations to deliver breakthroughs in AI and machine learning.\nThe cooperative research teams recently began collaborating with Intel and Dell Technologies in the development of new and enhanced radiology systems and techniques. Together, they built and trained a powerful new augmentative AI model, an AI radiologist to assist in detection of lung diseases.\nThe goal in this work was to aid radiology in delivering better and faster diagnosis of thoracic pathologies, such as pneumonia and emphysema, from chest X-rays. The researchers began the effort with a standard AI image recognition model called CheXNet, developed at Stanford University. This model offered state of the art machine detection of pneumonia. It leveraged a National Institutes of Health (NIH) dataset with over 112,000 multi-labeled chest X-rays spanning 14 different thoracic diseases.\nBy working with the HPC & AI Innovation Lab at Dell Technologies to scale-out the required compute and storage systems, the team was able to improve the performance of the AI radiologist, making it 187 times faster than before, thus reducing the model training time from months to hours. And diagnosis accuracy was bettered for most of the initial thoracic diseases represented in the dataset.\nBetter AI models, Better insights\nIf an altruistic goal of technology advancements is the betterment of humankind, then the application of augmentative AI and DL in Healthcare and Life Sciences is a brilliant example of how technology is truly enabling researchers, scientists, and practitioners to save time and lives.\nTo Learn More\nread the McGill University case study, watch the video, and listen to the podcast. Also, read the SURFsara case study and watch the video.\nExplore AI solutions from Dell Technologies and Intel.