Introduction
Machine Learning and Artificial Intelligence have been weaved into our lives in some form. Every newcomer to the field is trying to solve machine learning problems. But I have a different take on it. I think that one should not try to solve machine learning problems. And I have a reason to think so. Please read till the end to find the why.
Note: This article is just my sincere view of this field. Your views may differ. That is totally fine. And it would help if you surely shared your thoughts in the comment section.
The Motive to Think So
Newcomers to the field of Machine Learning and Artificial Intelligence are really enthusiastic about the problems. They try really hard to find machine learning problems. You may be asking – ‘So what? What is wrong with it?‘
Actually, there is nothing wrong with it. Even I tried to do the same when I started. But there is just one problem here. There are no real Machine Learning problems out there to be solved. They are just real-world problems that could not be tackled before. But are much more easily solved by using Machine Learning as a tool. And in fact, the implementation of machine learning in those areas is just one part. It can involve robotics, linguistics, philosophy, and other fields of engineering as well.
When I started out, I also searched for a few machine learning and artificial intelligence problems to solve. It was fun in the beginning. Then, even googling about problems did not turn up something specific. Slowly, I began to change my perspective. I looked at all the fields that implemented machine learning and artificial intelligence.
The revelation was really astonishing. I noticed that they are actually real-world problems where machine learning has been implemented. Even big companies are doing that.
For example, take the case of Tesla. They are not trying to build some AI-based vehicle. Rather they are trying to build safer transportation systems by automating driving while using artificial intelligence as one of the solutions. There are numerous other aspects that help to drive that car automatically. The cameras, light sensors, LiDAR sensors, and RADAR cover only some of them.
Another famous AI-based company, OpenAI, they are trying to create safe Artificial General Intelligence (AGI). At first glance, it may look like a particular problem. But go to their website, read their blogs, see their work on youtube. You will notice that they are actually trying to solve intelligence. That is a problem which has around for a long time and making scientists scratch their heads till now.
OpenAI’s work towards deep learning, trying to build intelligent game playing systems, all for one big thing. To solve intelligence and build AGI. The same can be said for any other big AI based company.
Conclusion
In the end, I just want to say that, don’t search for specific machine learning problems to solve. Try finding real-world problems where machine learning can be implemented. One of the best places to start is by reading research papers. They can give a really good perspective on what needs to be solved to make this world a better place. Again, all of this is just my own view. If your’s differ, surely let me know in the comment section. Follow me Twitter, Facebook, and LinkedIn to get regular updates.
