Whether a problem can be solved through machine learning is only a matter of how much data, creativity, and computational power does one have. Machine learning can be used to aid diagnosis, draw recommendations, classify stellar objects, protect animal life and tackle social issues. 

It can likewise be used to detect frauds, such as fraudulent credit card transactions and frauds in spreadsheets. Object , voice and face recognition, outlier detection and natural language processing (NLP) are few tasks that machine learning is recognized to perform very well already.

Sharp chatbots, able to pass Turing's test, can be built using machine learning. Brand new content, such as fake news (don't do that), sounds, music, poems, and samples can be all created by those models. You may find that It's very useful to generate samples from noise when you are short on data. These are only some of the numerous tasks that machine learning can smoothly handle.