There are many interesting topics right now, see the link:

https://www.techsparks.co.in/hot-topic-for-project-and-thesis-machine-learning/

• Feature learning from non stationary distributions (there are not even widely excepted benchmarks yet!).
• Complex regression problems (don't tell me about squared reconstruction error of MNIST or Tonronto Faces).
• Deep learning of heterogenuous data (most of the stuff is some sensory input).
• Multimodality in engineering domains (yes, there are ~3 papers on vision+text, but I am thinking more about sensory information from cars etc. sampled at highly different frequencies etc).
• Convincing work that MNAR works well with deep learning (not only in the vision domain).
• Robotics! Not the perception part, the control part: highly autocorrelated error models, provable guarantees, trajectory generation, real time constraints, non stationarity

Also refer following link:

https://www.quora.com/What-are-some-unsolved-problems-in-deep-machine-learning

Try biosignal classification, processing, and prediction that is gaining increasing importance in the biomedical research. There are a lot of very important problems that deserve work on them.

Each successful project can be applied in clinical praxis and save a lot of lives and lead to a better life for many people.

Some possible directions

- EEG classification & prediction

- ECG classification and prediction

- prediction of sugar & insulin levels in the blood of diabetic

- prediction of arrhythmias

- classification of the heart condition

- revealing hidden physiological mechanisms

and a lot more. A good idea is to contact some active biomedical researchers to find the whole set of starting information including EEG and ECG curves is time-consuming.

I have experience if being a consultant to an ML master degree student, it is a hard time to dive into physics, mathematics, and especially medicine quickly.

I offer consultancy to some future application in ECG studies including ML.

Jiří Kroc ....from where i will get the dataset for all these topics do u have any idea ?

One very good starting point is PhysioNet with its PhisoBank

https://www.physionet.org

I have been starting with it as well.

You can search for biomedical research groups that got a lot of data and do not know how to get a meaningful sense out of them.

Later, you will find a lot of data everywhere. The best is to collaborate with a group that produces data in experiments.

Good luck :-)

PS Do not hesitate to ask more. There is a lot of us working in the field and someone always help you.

Face recognition or pattern recognition concepts covers most of the machine learning concept

Start with something simple which is easy to grasp e.g. clustering then proceed to more sophisticated methods such as classification and artificial neural networks. Software packages have data sets included for different applications and methodd. I would recommend speech source localisation with deep learning as an interesting topic.

https://www.cs.waikato.ac.nz/ml/weka/

https://tel.archivesouvertes.fr/tel00913965/PDF/thesis.pdf

The best way to start with machine learning is to use Python with the SciPy library.

One course that will introduce you into the subject quickly is at

https://machinelearningmastery.com/machine-learning-in-python-step-by-step/

enjoy this awesome collection of ML and AI techniques. :-)

Refer the following websites(tutorials and projects in Machine Learning):

1. https://in.mathworks.com/campaigns/offers/machine-learning-with-matlab.html?gclid=Cj0KCQiAtbnjBRDBARIsAO3zDl8o8QoPpSERoKydTeOfc9aEzXPaByxl-1DtBdfRrjmuZGAXFg7X63MaAgoLEALw_wcB&ef_id=Cj0KCQiAtbnjBRDBARIsAO3zDl8o8QoPpSERoKydTeOfc9aEzXPaByxl-1DtBdfRrjmuZGAXFg7X63MaAgoLEALw_wcB:G:s&s_kwcid=AL!8664!3!182676450678!p!!g!!machine%20learning%20algorithms&s_eid=psn_40621491752&q=machine%20learning%20algorithms

2. https://www.ibm.com/account/reg/in-en/signup?formid=urx-30917&cm_mmc=Search_Google-_-Hybrid+Cloud_Data+Science-_-AS_IN-_-machine+learning+algorithms_Exact_&cm_mmca1=000026RL&cm_mmca2=10007643&cm_mmca3=&S_PKG=&cm_mmca7=1007811&cm_mmca8=kwd-15335911300&cm_mmca9=_k_Cj0KCQiAtbnjBRDBARIsAO3zDl_yjopt9vbHB_FYzPTk-BRzaZKDStYmSrMitju4lXUgncVyzG67SO8aAukzEALw_wcB_k_&cm_mmca10=332002948204&cm_mmca11=e&mkwid=_k_Cj0KCQiAtbnjBRDBARIsAO3zDl_yjopt9vbHB_FYzPTk-BRzaZKDStYmSrMitju4lXUgncVyzG67SO8aAukzEALw_wcB_k__898_26790&gclid=Cj0KCQiAtbnjBRDBARIsAO3zDl_yjopt9vbHB_FYzPTk-BRzaZKDStYmSrMitju4lXUgncVyzG67SO8aAukzEALw_wcB

3. https://in.mathworks.com/campaigns/offers/mastering-machine-learning-with-matlab.html?gclid=Cj0KCQiAtbnjBRDBARIsAO3zDl8UTdPmgcWd8BiRt2ykXfxzLkTXZMcZx8Z7TMwfR0-DQtA-edHpz_IaAu2bEALw_wcB&ef_id=Cj0KCQiAtbnjBRDBARIsAO3zDl8UTdPmgcWd8BiRt2ykXfxzLkTXZMcZx8Z7TMwfR0-DQtA-edHpz_IaAu2bEALw_wcB:G:s&s_kwcid=AL!8664!3!281794570925!p!!g!!machine%20learning%20tutorial&s_eid=psn_57384024392&q=machine%20learning%20tutorial

4. https://www.udemy.com/machinelearning/?gclid=Cj0KCQiAtbnjBRDBARIsAO3zDl_IkCahnP2wCAZsVmI2u8jAwEShMMQK4OrFfM2sBxi_hqj1iJZ2TIAaApvxEALw_wcB&k_clickid=514a405d-98fd-4c09-88db-66365c7b8871_k_Cj0KCQiAtbnjBRDBARIsAO3zDl_IkCahnP2wCAZsVmI2u8jAwEShMMQK4OrFfM2sBxi_hqj1iJZ2TIAaApvxEALw_wcB_k_&matchtype=e&utm_campaign=NEW-AW-PROS-ROW-PROF-Business-MachineLearn-EN-INDIA_._ci_950390_._sl_ENG_._vi_PROF_._sd__._la_EN_._&utm_content=_._ag_tutorial_._ad_263695139506_._de_c_._dm__._lo_9075363_._&utm_medium=udemyads&utm_source=adwords-row&utm_term=_._pl__._pd__._ti_kwd-294938473150_._kw_machine+learning+tutorial_._

5. https://www.simplilearn.com/artificial-intelligence-masters-program-training-course?utm_source=google&utm_medium=cpc&utm_term=%2Bmachine%20%2Blearning%20%2Bproject&utm_content=311611388045&utm_campaign=Search-AIML-Machine-NA-NA-IN-Main-Masters-Desktop&utm_adgroup=AIML-Machinelearning-Project-BMM-New&utm_campaignid=1628235961&utm_device=c&utm_network=g&gclid=Cj0KCQiAtbnjBRDBARIsAO3zDl-F4fdfkOE9DgDvEmXQRuJ6xd7QwCc_s3Kmm9IYyq9tcrNqibf54FIaAsTLEALw_wcB

6. https://www.dezyre.com/article/top-10-machine-learning-projects-for-beginners/397

IBM course is useful for begineers. Simplelearn eventhough costlier it will be useful. But anyway u have to learn the following two software languages:

1. Matlab. very efficient tool for Deep learning and Machine learning.

2. Python. Those who are want built new packages at free of cost, they may prefer Python software