Ram D. Sriram – författare

Realizing the promise of digital technology will depend on the ability to share information across time and space from multiple devices, sources, systems, and organizations. The major barrier to progress is not technical; rather, it is in the failure of organizational demand and purchasing requirements. In contrast to many other industries, the purchasers of health care technologies have not marshaled their purchasing power to drive interoperability as a key requirement. Better procurement practices, supported by compatible interoperability platforms and architecture, will allow for better, safer patient care; reduced administrative workload for clinicians; protection from cybersecurity attacks; and significant financial savings across multiple markets.

With funding support from the Gordon and Betty Moore Foundation, this National Academy of Medicine Special Publication represents a multi-stakeholder exploration of the path toward achieving large-scale interoperability through strategic acquisition of health information technology solutions and devices. In this publication, data exchanges over three environments are identified as critical to achieving interoperability: facility-to-facility (macro-tier); intra-facility (meso-tier); and at point-of-care (micro-tier). The publication further identifies the key characteristics of information exchange involved in health and health care, the nature of the requirements for functional interoperability in care processes, the mapping of those requirements into prevailing contracting practices, the specification of the steps necessary to achieve system-wide interoperability, and the proposal of a roadmap for using procurement specifications to engage those steps. The publication concludes with a series of checklists to be used by health care organizations and other stakeholders to accelerate progress in achieving system-wide interoperability.

This book surveys the state of the art in explainable and interpretable reinforcement learning (RL) as relevant for robotics. While RL in general has grown in popularity and been applied to increasingly complex problems, several challenges have impeded the real-world adoption of RL algorithms for robotics and related areas. These include difficulties in preventing safety constraints from being violated and the issues faced by systems operators who desire explainable policies and actions. Robotics applications present a unique set of considerations and result in  a number of opportunities related to their physical, real-world sensory input and interactions.