Welcome to the BIOMEDigital session viewer. Browse and sort the growing list of sessions by day or track. All sessions are shown in Eastern Time.
Date: Thursday, November 5
Time: 2:00pm - 3:00pm
Track: Center Stage
Vault Recording: TBD
As Bayesian Artificial Intelligence has progressed, there has been growing intrigue in the use of AI in Medical Devices. There are certainly challenges when it comes to the application of Artificial Intelligence in the real world. The traditional approach would be to train a machine learning or deep learning system and then feed it data to uncover patterns that can then be used to alter designs or performance parameters.
The traditional approach to AI deployment would not lend itself useful in all applications and real world situations. In many cases, the optimization of parameters would have to be applied on the spot for maximum effectiveness. This is where edge computing comes into play. Data is collected at the point-of-use such as on the manufacturing floor or the point-of-care and using efficient edge computing resources such as hardware and algorithms, a partial or full-blown analysis is performed on the spot and the results are used to make dosage decisions such as insulin or drug delivery, treatment parameter changes such as in energy delivery, surgical treatments and others. Edge Computing can also have unique applications in helping prevent addiction, providing cost-savings and the reduction of side-effects, medical errors as well as other advantages.
The presentation will focus on how to use low-cost, low-profile devices in combination with algorithms to optimize for cost, performance, energy efficiency and other factors.Hardware and Software prototyping resource examples will be provided. The takeaways will include an understanding of current and future opportunities and limitations of edge computing.