AI

Artificial Intelligence for Anaesthesia

Nothing too fancy on the AI front, we are mostly using standard backprop, fully feed-forward Artificial Neural Networks (ANNs). The core GPU-accelerated Neural Networks are from the excellent open-source brain.js. The project design, training data, neural-network structure and parameters, implementation, interface, training, detailed audit, and ongoing maintenance are of course entirely by JET5. We briefly outline of each of our projects here, including links to development pages and prototypes.

These Expert Systems are being developed in collaboration with the clinical staff involved and are subject to ongoing audit. They are not currently available for use outside the QEUH in Glasgow. Everything here is strictly for information and interest only.

    Suitability for Dental Sedation

An expert system designed to inform the existing pathway from "patient assessment" to "anaesthetist-delivered IV sedation" for paediatric conservative dentistry. This service is currently seeing an increase in referrals from a variety of sources, and we are keen to maintain our high success rate in this challenging sub-specialty.

    Pre-op Assessment

At the QEUH this consultant-led clinical service is under stress with ever increasing workload and limited resources. Another AI-based expert system which complements the current assessment process and will hopefully reduce the workload.

    Pain Management

Another consultant-led clinical service which is experiencing increasing workload in an environment of limited resources. Encapsulating consultant-level expertise in an easy-to-use AI Expert System will enhance our ability to provide and expand this high-quality service.