This site uses cookies to deliver our services and to show you relevant ads and job listings. By using our site, you acknowledge that you have read and understand our Cookie Policy, Privacy Policy, and our Terms & Conditions. Your use of Stack Overflow’s Products and Services, including the Stack Overflow Network, is subject to these policies and terms. I Accept

Dr Ada - Medical AI

ORBAI - Dr Ada: Medical AI

We need a healer that can take the wonders of North American and European medicine, its power to diagnose and treat disease, encapsulate it in an AI and take it everywhere, to the rest of the world, to heal the sick, and get them on their feet so a better world for them is even possible. Introducing an advanced medical AI with the same diagnoses and treatment capability, using the advanced medical databases…for the whole world, Dr Ada, Medical AI

(click below for video)

Have you ever been to the Doctor and tried to describe what was wrong with you, and you just didn’t seem to be able to communicate it nor get through to them, and you ended up being sent home without a diagnosis, or worse, being misdiagnosed?

Dr Ada is a 3D interactive character on mobile, who can interview you about your medical and physical background, lifestyle, and then about your current symptoms. She can take your basic vitals, like blood pressure, pulse, temperature, ... using commercial bluetooth sensors  and then search using this information, to see what medical conditions may be causing your symptoms. She can create a detailed medical report, with historical data over the past weeks, and pass it on to your doctor, making it more likely they will understand and diagnose you correctly.

If no doctor is available, Dr Ada can give you suggestions of what the condition might be and suggest treatment. If you are home and sick, she can monitor your vitals and symptoms every day, and compile a report with graphs for the week.

On the back-end, all the diagnostic data – interview, symptoms, vitals, medical imaging, lab results, … can be encoded and combined into a snapshot of the patient’s medical state at a given point in time. As the patient undergoes treatment, these tests would be repeated to track the progress, creating a series of snapshots tracking the patient’s medical condition over time, that forms a narrative, a story of the progress of their disease, whether it is a flu for a few weeks, cancer over months, or diabetes over years. Because our Medical AI is more than one instance, and is an AGI - she is compiling these medical narratives for every one of the millions of patients she treats every day, and they can be used by the Medical AI to learn about the progression of different human diseases in great detail, and to use all this knowledge to forecast the patient’s medical condition into the future.

However, we need to track more than the patient’s medical state. We also need to track the treatment regimen, and what the doctor is doing to change the patient’s medical state. We can create snapshots of the treatment, encoding the medications, dosages, frequency, and any procedures or therapy that are done. The patient can even record their exercise or diet progress.

Now, the narrative for a patient’s medical story is more complete – we have a narrative chain consisting of alternating evaluation and treatment engrams that track the details of their medical condition and what was done about it in great detail, encoded in a way that we can train a predictor on it, or make an adversarial model where the AI can run multiple simulated chains of the patient’s health through time, and change the treatments, dosages, intensities at each point to see if it can come up with an optimal treatment plan for that patient. This is very powerful, as most doctors resort to doing this by guessing, doing trial and error with their patients, where this would be an AI tool that could look months, or even years ahead to plan a more effective treatment plan that would have a better future outcome. It could prescribe treatments preemptively to keep the person from even getting ill years from now, based on detecting subtle early warning signs of a condition early.

Check out my Tech Specs.

Also, by doing this with millions of patients, and autoencoding the progression timelines laterally, the medical AI will build up an enormous knowledge base of human health, disease, and the progressions of those diseases and the treatments that work best. When this AI is distributed globally, it can learn the medical practices and treatments of other cultures. Maybe a doctor in California would prescribe a series of radiation treatments to shrink a specific type of cancer tumor, but a rural doctor in Central America knows that the excretion of a specific insect, when extracted and modified by mixing it with lye and cooking it – is a far more effective treatment that destroys the tumors completely.

In addition to a medical treatment AI, our global medical AI would be the world’s largest scale pharmaceutical research platform – building up these databases of medications, their efficacy, and the direct results on patients over time, and by allowing pharma companies to mine the database and use the AI in clinical trials, we have a whole other application for a customer with very deep pockets.