There are a lot of problems in medicine that needs the newest technologies in computer science fields like Machine Learning and Deep Learning to solve them.
Can anyone mention some of these problems that are unsolved till now?
Let's start with a field that I know the best: biosignals processing, clarification, and prediction. There are many unresolved problems like
* Prediction of arrhythmias
* Prediction of epileptic seizures
* Reconstruction of physiological interdependencies of various physiological processes using biosignals.
* Reconstruction of physiological paths in the brain from biosignals.
* Assessing the health condition of people using biosignals.
We can do the following research
* Data mining for disease-genome dependencies from available databases.
* Prediction of disease spreads using internet searches.
* Assessment of the population health using internet searches.
* Cross AI methods with deep knowledge of complex systems theory and apply it to medicine -- this is my area of research.
* Study and predict drug interactions and side effects in patients. This will save a lot of unnecessary suffering in those using medical drugs.
* The above can be supported by an active search through all available data for possible, future drug interactions prior to their application to patients.
* Such research can help medical doctors to avoid deadly or highly damaging drug interactions. Each patient reacts differently to the same drugs! We need to know why and especially when it happens!
* Start development of advanced AI methods tailored towards the needs of bio-medicine.
All the above depends on how reliable databases of biosignals, medical records, bio-imaging, laboratory results, and many other database build.
When you want to have successful research in the field of AI, perfect databases that are open-access are a must. We have an extreme shortage of those databases. You can build a very successful carrier by building such a database(s). :-)
This is just a short list of all possibilities. :-)
Let's start with a field that I know the best: biosignals processing, clarification, and prediction. There are many unresolved problems like
* Prediction of arrhythmias
* Prediction of epileptic seizures
* Reconstruction of physiological interdependencies of various physiological processes using biosignals.
* Reconstruction of physiological paths in the brain from biosignals.
* Assessing the health condition of people using biosignals.
We can do the following research
* Data mining for disease-genome dependencies from available databases.
* Prediction of disease spreads using internet searches.
* Assessment of the population health using internet searches.
* Cross AI methods with deep knowledge of complex systems theory and apply it to medicine -- this is my area of research.
* Study and predict drug interactions and side effects in patients. This will save a lot of unnecessary suffering in those using medical drugs.
* The above can be supported by an active search through all available data for possible, future drug interactions prior to their application to patients.
* Such research can help medical doctors to avoid deadly or highly damaging drug interactions. Each patient reacts differently to the same drugs! We need to know why and especially when it happens!
* Start development of advanced AI methods tailored towards the needs of bio-medicine.
All the above depends on how reliable databases of biosignals, medical records, bio-imaging, laboratory results, and many other database build.
When you want to have successful research in the field of AI, perfect databases that are open-access are a must. We have an extreme shortage of those databases. You can build a very successful carrier by building such a database(s). :-)
This is just a short list of all possibilities. :-)
This question needs further specifation. You could also ask
What are the unsolved questions in physics?
What are the unsolved questions in biochemistry?
Etc
etc
Very broad question.
It sounds like the question at the very beginning of sequencing era "What are unsolved questions in genetics?". We all see the answer for that one. After many years of sequencing, there are lots of unsolved questions.
On the other hand,machine learning could need a training dataset. This way, the machine could provide some clarification for an already existing hypothesis.
AI and all related technologies are powerful tools. but can't just solve everything.
Shahab Pasha Iman Hassan Ibrahim Jos Wielders Jiří Kroc Thanka a lot for your replies. I really appreciate that!
- Badges
- Science topic
- Similar topics
- Medicine