After learning about some examples of AI applications in the diagnosis phase and in the development of new tools, let's talk about AI and predicting the future. Everyone dreams of the ability to predict the future. Such an ability could avoid some problems in our daily life.
 |
| Data Analysis - Photo by Carlos Muza on Unsplash |
Could AI leverage us to reach that desired dream? We all heard about predicting AI algorithms applied especially on our economy. However, the big question here is: Can we use AI to predict diseases, new diseases, or even pandemics? Do we already have this kind of algorithms?
Google's Scientists developed an AI system capable of analyzing scans of the back of an eye so it can predict data about the patients [1]. Some of the predicted data is related to the person's age, blood pressure, or if he/she smokes. Such indicators are directly related to cardiac events, like heart attacks. Thus, using this method, we do not need blood tests to get the necessary information for a possible heart attack diagnosis. Therefore, we have a quicker evaluation of cardiac event's risk. Moreover, the implemented AI system has approximately the same accuracy percentage as the current leading methods.
Some other companies are already applying AI algorithms to analyse health care reports. Some approaches use Natural language Processing (NLP) to identify severe diseases on the reports. Then, based on those findings, the predictive algorithms can forecast the impact of these diseases, or in cases of infections diseases, simulate the spread of it around the world. BlueDot is one of the companies investing in this field, trying to predict pandemics and alert soon enough to make the difference.
Indeed, this company alerted for the well-known COVID-19 pandemic. They also predicted some of the most attacked countries, like China, USA, and Italy. However, they are still struggling with their AI implementations because the algorithms are becoming less precise as the virus is spreading around the world. One of the biggest causes of it is the lack of well-labelled symptoms, or correctly know how the virus propagates from human to human. We still do not have the necessary information about the virus to feed the AI algorithms with reliable data. Nevertheless, we still can use these algorithms, taking into account some margin of error, to timely prepare the next city being predicted as a disease's focus.
I think a good AI use on this COVID-19 pandemic could be the risk prediction for every person. What about an AI algorithm that analyses our medical history and outputs if we are in the danger zone in case of getting the virus? It could tell us if we will have severe or mild symptoms. Such information could leverage the isolation of the ones with the highest risk of severe symptoms and curve down the death line on our charts. The system could also alert a non-risky person if he/she lives with someone who could suffer from the virus. A system like this helps everyone's health and even economics because only the ones with higher risks (and their closest family members) would be isolated. I hope a system like this will be out in the next years. Unfortunately, to fight COVID-19, we are still blind on this level (at least until the moment I am writing this post). Only after being infected with the virus, we can know how our system handles it.
Do you have a new idea using AI to predict diseases or epidemics? Please let me know.
References:
- Poplin, R., Varadarajan, A. V., Blumer, K., Liu, Y., McConnell, M. V., Corrado, G. S., … Webster, D. R. (2018). Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning. Nature Biomedical Engineering, 2(3), 158–164. https://doi.org/10.1038/s41551-018-0195-0
Comments
Post a Comment