Bayesian Network AI Medical Diagnostic Tricorder

In 2019 I proposed a design for a Medical Diagnostic Tricorder (Star Trek, Gene Roddenberry) device for real-time in Theatre (at time of war) Triage and Treatment. The device could supplement medical diagnostics by Doctors or Surgeons or be used by anyone without medical training if required. This would then make the human survival rate extremely high with limited, still functioning and dispersed skilled clinicians.

Making Decisions with Unknown Factors

Bayesian Network is a probabilistic graphical model for representing knowledge about an uncertain domain where each node corresponds to a random variable and each edge represents the conditional probability for the corresponding random variables. Essentially the more data sources that are brought to bear the higher the probability of a correct assessment. You can find out more about Bayesian Networks on Wikipedia.

Primary Scenarios of Use

The primary scenarios for use where post a Bio or Nuclear weapon usage where communications would be limited or non existent. Although these are extreme scenarios it was the focus of the potential client, however the device could just as easily be utilised in locations with financial or medical poverty.

The device would therefore be constructed as an intrinsically safe device to CAT1/2 and IP68 (dust tight and protected from full water immersion) to avoid contamination and utilise an augmented intelligence algorithm repurposed from an existing AI used in threat diagnostics with the military and security services. The augmented intelligence algorithm is built using thin client architecture making it capable of being self contained in a portable device.

The complexities around Bio or Nuclear weapon identification and treatment are extended due to patient medical history, clinician skills and prior knowledge, access to medical equipment and the physical operating environment. The intention was that the unit would be portable (easily carried and used by one person), secure, replaceable at low cost (relative low cost) and mass producible. This evolution on the existing augmented intelligence algorithm has two aspects a new query and a new container.

This solution utilises the exact same process that a medical professional would to diagnose health issues so would both establish, reenforce standard practices and drive the evolution of faster and potentially less qualified clinicians to be able to make decisions on patient triage and care.

Other Proposed Usages

Other proposed usages of this technology includes linking large-scale longitudinal data on health and treatments with genomic and biomedical analyses to develop new technologies to improve patient outcomes. This combined with enhanced (though augmented intelligence algorithm) enables the development of new products and services such as machine learning based data analytics tools that will help diagnose diseases earlier and assist clinicians the best treatment for individual patients.

How will it work?

The full details are obviously held under IPR however the model below was used to describe how the augmented intelligence algorithm would support Banking and other Financial Service organisations in creating bespoke products and services for customers. Additionally these concepts were used to define the Continuous Listening, Improvement and Optimization AI component of the Customer Agility Framework.

Last Word

The purpose of this article is to show the Design, Engineering and Scientific knowledge of Karl A L Smith, if you'd like to collaborate with him, please contact him, he's 'always open to the conversation' is speaking about yourself as a third person a sign of madness or genius. I shall discuss this later with myself lol!