ML is just a tool. Your question is basically " I want to do a PhD in Screwdriving, but can people suggest me what I could use the screwdriver for ?"

What "problem" are you trying to solve, which makes a reason for you to start on a PhD ?

When you want to do important biomedical research, I recommend you to use ML & AI in studies of physiology, ECG recordings, EEG recordings, gait studies, emotion studies, and a lot more what is going to become mainstream research in medicine in the coming decade.

There is a lot of space to find ways how to predict health development from acquired electronic data by using different medical devices and wearable devices. Approaches that are currently being in development will sooner or later lead to personalized, predictive medicine enabling to resolve health issues even before they will reach the stadium of clinical symptoms.

You can work on technology that we currently know only from sci-fi movies and literature. On top of it, you can be very close to the development of complex systems theory, which already becomes the mathematics of the 21st century. You will be in the finest possible company while working there.

Be very selective in choosing your field as there are present big problems with acquiring data in biology and medicine. Data are usually produced by people who do not understand the core of ML & AI and expect that any approach of their creation is OK. It is more complicated than it seems to be from the first sight. You can work a year or two to find that data is inconsistent, corrupted, and most probably insufficient to make reasonable conclusions.

Good luck with your choice :-)

Jiří Kroc Actually, the EEG Recording seems to be a very interesting idea.They are mainly used at sleep researches, Thank you so much for helping me choose this, and I was wondering if you have more informations about the EEG recordings, or if you know any platforms where people share the approches that have been made using the EEG recordings.

I do not have personal experience with those EEG databases but there is one database of physiological data called Physio Net

https://www.physionet.org

It would be a good idea to start there. There are a couple of good areas where I know research results of EEG processing, which are having very good results but data are probably there. It is not sure for me.

In general, data in physiology and medicine, are very difficult to acquire as researchers usually keep them for themselves and request coauthorship in the case of their sharing.

This is the problem that should be addressed in the future as it brings into medicine unhealthy monopolization data paid from public resources. In any case, it is a story for a completely different question.

Good luck in your research and do not hesitate to ask again. There are a lot of researchers who know the answer. :-)

Indeed, EEG is very promising direction to choose . Beside physionet, where you can find some, but not that huge resources, you can try this

ieeg.org repository

http://www.ieeg.org/

Just to add, you can use EEG in several important problems solving like the depth of anesthesia, the sleep stages (and sleep mechanisms in general), predicting epileptic seizure (and confirm epileptic focii). lear about the developmental processes, short-term memory formation, learning, detecting depression (and bipolar depression, schizophrenia, etc) and predict the responders/outcome of the treatment, predict stroke, study AD, PD, Dementia...the list is long

Completely agree with a colleague Dr. Cukic. I have studied those papers dealing with the prediction of anesthesia depth, sleep stages, prediction of epileptic seizures, etc. from ECGs. They are very good and they use up to date knowledge of complex system measures based on entropy in combination with ML techniques.

Personally, I work on EEG and prediction of arrhythmias using quite similar techniques. I got very good results with those methods. It is possible to predict arrhythmias from ECG recordings at least hours prior to their onset. It is an incredibly beautiful research area that has great potential.

In general, predictive medicine is arriving, we are developing new techniques that will change medicine from a reactive one towards proactive, preventive, and personalized one. You can join us :-)

Milena B. Čukić Thank you so much for giving me these ideas.First, of all I'm going to ckeck the availability of Data, and try to make some simple predictions using a simple Artifical Neural Network.

Jiří Kroc Do you suggest that I use more then an algorithm in such an approch? and what do you think if I go for the Artifical Neural Network?

ML is just a tool. Your question is basically " I want to do a PhD in Screwdriving, but can people suggest me what I could use the screwdriver for ?"

What "problem" are you trying to solve, which makes a reason for you to start on a PhD ?

Mathieu Clerté yes you are right, I know everything about the tool that I want to use, which is the ML, but I'm really confused which problem to solve. I have in mind some ideas but don't know what to choose.

If you have any ideas I would so glad to hear them.

Imane Habib Msamri But is your goal to do a thesis on the tool itself, showcasing it with some practical example ? Or is the goal to use the tool in your particular research field (What is it by the way ?) ?

Dear Friends and Colleagues of RG

I propose the use of various advanced information processing technologies for the purpose of improving learning processes as part of learning machines.

The development of information processing technology in the era of the current technological revolution defined by Industry 4.0 is determined by the application of new information technologies in the field of applications in technical industries.

The current technological revolution known as Industry 4.0 is motivated by the development of the following factors: Big Data database technologies, cloud computing, Internet of Things, artificial intelligence, Business Intelligence and other advanced data mining technologies.

The use of advanced information processing technology for the improvement of learning processes as part of machine learning would allow for the continuation of technical progress in this field.

Best wishes

One of the most interesting fields for machine learning application is structural health monitoring/condition monitoring, among others. Statistical pattern recognition has a potential to aid the efforts in increasing the safety and reliability of critical structures - bridges, tunnels, dams, skyscrapers, as well as in aircraft and automotive sectors.

This is a very interested topic and excellent work has bee done by the researcher.

I propose research in the field of using Machine Learning technology to improve online robots, avatars, which act as remote advisors and are installed on the websites of public institutions, offices, internet technology companies, banks, other financial institutions, etc.

Regards

Dariusz Prokopowicz

I feel curious about what field you chose in your thesis because I have same problem.I'm really confused

Here you have two sources

https://medium.com/@ODSC/10-compelling-machine-learning-dissertations-from-ph-d-students-c85a2d73a7e

https://www.ml.cmu.edu/research/phd-dissertations.html