It depends on why you're researching on machine learning! What's the purpose? What are you really trying to find out?
Is it for your school project?
Has your college professor given you the assignment on, say, "how will machine learning change the world?"
Or, is it your boss at the office who has asked you to do research on how the company can power-up its marketing strategy using machine learning?
The scope of machine learning is extremely wide, playing a significant role across major industries, right FROM healthcare, manufacturing, and cybersecurity TO e-commerce, banking, and transportation.
So, before you start researching on machine learning, it's essential that you establish what you’re trying to find out, what's the purpose, what problems' solutions you’re looking for, and so forth.
And then, of course, if you’re researching on machine learning just for leisure purpose, there's literally nothing you should consider. You can begin with anything and proceed per your preference and interest.
All said, remember, when you’re doing your research on any topic:
• Consume contents only on renowned and reliable websites. Avoid shady platforms.
• Read articles in full before coming to the conclusive point. Half-baked knowledge is dangerous.
• Don’t just read headlines and move on. We're in living in the 'headline society' where the number of explosive, misleading and controversial headlines, even on reputed platforms, is on the rise.
• Most importantly, give yourself enough time. If the research is for assignment, start as early as possible.
Hope this answers your question.