Solving Fakenews problem in Nigeria with the use of Machine Learning/deeplearning/AI : Anyone, Institution, brand

You asked for it. Now you have it

Machine Learning is a buzz word. ‘Learning’ is a self explanatory but Machine may not be. A car is a machine because its movement is faster than humans or animals. A car or bike has other functionalities that makes it more efficient and effective than humans but guess what, humans built and programmed every computer to engage in more powerful activities. Machine Learning/deeplearning/AI can ve used interchangeably here to avoid too much complexity/ambiguity.

It is no longer news that internet, smartphone, connected devices are already revolutionizing the way we interact with each other. Communication is faster and almost irretrievable and so what goes out, goes out and seen by many in different part of the world. Almost irretrievable!

“According to Forbes, IN A MINUTE, 98,000 tweets are shared on Twitter, about 600 videos are uploaded on YouTube, and about 1500 blogs posted on several blogging platforms, all over the world.”

Now let us simplify the definition of Machine Learning. Machine Learning  in the context of Fake news is a set of computer programmes that learns from data by finding patterns and then attempts to make some analysis, with the objective of making predictions. So much so that when it sees future occurrence, it can easily detect it and flag it off (in this case whether the news is fake or not )

What is fake news. Fake news is like ‘April fool’, Fake news is Misinformation. It could come with a badly captioned headline, someone’s opinion that is false. It’s done in such a way that the false content is intended to mislead people. Let me quickly say that people have 100% right to convey their opinions e.g ‘ I don’t like xy politician or xy celebrity’ They have the right to say what, who, why they don’t like what they like.

However, it is fake news to say ‘xy is dead  or ‘xy is a smuggler’ when in the real sense he is not. Fakenews come from un-reliable sources, or from un-verified humans and intended to pull one down

Machine Learning, deep learning, AI (all inter-related) seek to find patters from past event and algorithms are plugged to separate 1 and 0 (in this case False and True)

Let me also quickly add that detecting fake news is no easy task as the first steps requires some ‘dirty work and hard work’ of collating and properly labelling which news is fake or not. It is with these labels that a machine learns from (remember the popular saying ‘gabage in, gabage out’).  A lot more also goes into this dirty work and only a Machine Learning Expert can guide and run you through this project. With every hardwork, there is a victory

In conclusion, now we know some basics, properly verifying information as an individual, journalist/ media house is extremely important in curbing fakenews.

Please note: This write up is not a comprehensive tutorial for machine learning, deeplearning, AI but for those who  do not know and those who have asked the question on how ML can help curb fake news. I have tried to simplify it and not make it ambiguous or boring  for these new set of  readers.

For the readers who want to engage me further in a fakenews detection project or more learning, please see my contact below

Chioma Omolaye

[email protected]

BigML Engineer

Founder, TechEduHub