Scientist have created an incredibly accurate AI that predicts how you die

Would you want to know how you will die?

Would it help you take measures to prevent that death, or would it just sap your carefree youth in which you live ‘in the moment’ because you feel invincible?

If you knew you would die from liver failure you probably wouldn’t chug 4 Lokos in the parking lot with your friends (which we already know is extremely dangerous in and of itself). Or maybe you would – there’s plenty of people who drink even though their families have a genetic predisposition towards alcoholism, or who try a cigarette after watching a loved one die of cancer.

I suppose the point is that if you knew, you’d have the power to do something about it. Or not. YOLO.

We all know that experts have created prediction models in the past to some success. The Cox Model of prediction simply takes into account age and gender. This method is the simplest and not very accurate, though it is still used. A multivariate Cox regression model takes more variables into account, such as genotypes, but the degree in which it can predict death is still of nominal value.

And, like always, our machine overlords do a better job of this. My new saying is this: Nothing is certain except death, taxes, and the superior rise of AI over the human race.

PLOS ONE, an online publication that features peer reviewed scientific research, published some controversial data from a team of scientists at the University of Nottingham. They found they could predict death in 80% of the 14,500 cases who ended up dying during their study. Now of course it can’t tell you if you are going to die due to a strike from an Apollo Asteroid, or if you will be eaten by zombies. But for a large number of frequent killers like cancer or stroke, it can give an almost eerie-like insight.

A lot of prediction applications on death take one or very few factors into account. Calculating the risk of someone dying only from heart disease is nowhere near as complex as predicting their risk of death with all of possible culprits of premature death.

Get our ‘Everything hurts and I’m dying’ shirt, tank, hoodie, and more, on sale now in our store!

These complicated calculations take into account a wide range of demographic, biometric, clinical and lifestyle factors for all 502,628 participants aged between 40 and 69. They even assess their dietary consumption of fruit, vegetables and meat per day. The study began in 2006 and ran to 2010, with a follow-up in 2016. All of these extremely individualistic characteristics can be compiled into their algorithm and give us a good look into possible future concerns.

PLOS ONE goes into detail about their methods here:

The AI machine learning models used in the new study are known as ‘random forest’ and ‘deep learning’. These were pitched against the traditionally-used ‘Cox regression’ prediction model based on age and gender – found to be the least accurate at predicting mortality – and also a multivariate Cox model which worked better but tended to over-predict risk.

Professor Joe Kai, one of the clinical academics working on the project, said: “There is currently intense interest in the potential to use ‘AI’ or ‘machine-learning’ to better predict health outcomes. In some situations we may find it helps, in others it may not. In this particular case, we have shown that with careful tuning, these algorithms can usefully improve prediction.

“These techniques can be new to many in health research, and difficult to follow. We believe that by clearly reporting these methods in a transparent way, this could help with scientific verification and future development of this exciting field for health care.”

This study is an extension on previous work by the team at the University of Nottingham, which showed that four different AI algorithms, ‘random forest’, ‘logistic regression’, ‘gradient boosting’ and ‘neural networks’, could better predict issues of the heart than algorithms used in current cardiology guidelines across the world.

PLOS ONE tells us the findings of this earlier study:

14,418 deaths (2.9%) occurred over a total follow-up time of 3,508,454 person-years. A simple age and gender Cox model was the least predictive (AUC 0.689, 95% CI 0.681–0.699). A multivariate Cox regression model significantly improved discrimination by 6.2% (AUC 0.751, 95% CI 0.748–0.767). The application of machine-learning algorithms further improved discrimination by 3.2% using random forest (AUC 0.783, 95% CI 0.776–0.791) and 3.9% using deep learning (AUC 0.790, 95% CI 0.783–0.797). These ML algorithms improved discrimination by 9.4% and 10.1% respectively from a simple age and gender Cox regression model. Random forest and deep learning achieved similar levels of discrimination with no significant difference. Machine-learning algorithms were well-calibrated, while Cox regression models consistently over-predicted risk.

We’re seeing the application of AI predictive models that could help in a variety of ways. Google has created an AI that can predict things like how long people may stay in hospitals, their odds of re-admission and, of course, their chances of death.

Bloomberg states that Google’s AI can

sift through data previously out of reach: notes buried in PDFs or scribbled on old charts. The neural net gobbled up all this unruly information then spat out predictions. And it did it far faster and more accurately than existing techniques. Google’s system even showed which records led it to conclusions.

These amazing strides in AI could revolutionize the healthcare industry. Perhaps in the future we can ensure small mistakes which lead to serious outcomes can be avoided, that the running of our hospitals can become more efficient, and help staff spend more time on patients instead of paperwork.

So hopefully in this age of AI, where fears of robot masters are creeping behind every victory in AI, robotics, and science, we can calm our anxiety and focus on the positive application of these topics.

https://twitter.com/CromarBlake/status/1112881571118415874?ref_src=twsrc%5Etfw

I constantly say "please" and "thank you" to #Siri in hopes that, one day when our #robotoverlords take over, that they will remember I was benevolent.

— Angie (@ABoBangie) March 20, 2019