Gastroenterology and Artificial Intelligence: 3rd Annual Artificial Intelligence Summit (On-Demand) | September 2021-9 - AI in Healthcare Current Advances in AI and How it Translates to Medical Applications

9 - AI in Healthcare Current Advances in AI and How it Translates to Medical Applications

Pdf Summary

The document discusses the use of artificial intelligence (AI) in healthcare, specifically in the areas of smart hospitals and robotic surgery assessment. In smart hospitals, AI is utilized for tasks such as body temperature screening, safe social distancing, fall prevention, contactless controls, and patient monitoring. The Da Vinci Surgical System is highlighted as a premier surgical robotics platform with three main components: a vision system, surgical cart, and surgeon's console. The document emphasizes the importance of objective skills assessment in robotic surgery, as human-to-human assessment can be subjective. Skills assessment in robotic surgery is related to patient outcomes, and the classification of suture gestures is identified as a useful method for assessing surgeon skill levels. A dataset of labeled surgical videos is mentioned, which includes actions such as needle positioning, targeting, driving, and repositioning. The document presents a Convolutional Neural Network (ConvNet) architecture for feature extraction, fusion, attention modeling, temporal processing, and skill assessment. The results show that AI can achieve state-of-the-art skill assessment performance with 95% AUC.<br /><br />Furthermore, the document introduces NVIDIA's Clara Application Frameworks for healthcare, including Clara Imaging, Clara Discovery, Clara Guardian, and Clara Parabricks. Clara Imaging is an end-to-end AI application framework for development and deployment, offering pre-trained models, labeling data, and training capabilities. The concept of federated learning is discussed as a privacy-preserving and extensible collaborative learning approach. Differential privacy and homomorphic encryption are mentioned as methods to prevent data leakage and preserve privacy during collaborative learning. Clara Guardian is presented as a solution for smart hospitals, providing speech models, pose estimation models, heart rate estimation, gesture models, and sample apps. The use of AI in computational drug discovery is also highlighted, with applications in literature analysis, real-world data, chemical compounds, and imaging. Additionally, the document mentions AI4QuantumChemistry, a computational platform for faster molecular orbital calculations in quantum chemistry. Overall, the document showcases the potential of AI in revolutionizing healthcare through various applications and frameworks.

Keywords

artificial intelligence

healthcare

smart hospitals

robotic surgery

skills assessment

Convolutional Neural Network

Clara Application Frameworks

federated learning

computational drug discovery