Talking about changing the world sounds like a cliche to a lot of people, but I do it anyway. And for me, it's not just about changing the world, it's about becoming a better human and reaching my highest potential. So when I decided to quit my job back in January 2019, everyone including my parents thought I was crazy. I did not have a plan B. I just knew that I wanted to learn Machine Learning, and I was ready to go anywhere to do that. I was obsessed with applying ML to real-world problems and I just had to pursue this field with all my heart and soul. That's how I ended up in Canberra studying my Master's in Machine Learning and Computer Vision at ANU.
Now incidentally, the same year, Omdena had started its projects where they partner with NGOs, companies and government organizations to solve problems for them through data analysis. I was very surprised to see people volunteering their time for Omdena projects without expecting anything in return. That is how my inkling took a very concrete form - using Machine Learning to do proof of concept models in order to solve "wicked" problems.
Wicked problems are such intertwined, multi-disciplinary problems, that solving them through traditional and one-track approach is almost impossible. When a large number of people from various disciplines come together to solve such problems, then there is some hope of getting some solution. I was always interested in such wicked problems - hunger, poverty, farmer suicides are all the issues that sometimes kept me up at night and I am glad to see such problems are starting to get tackled now. Do you want to know how? Read on.
Let's talk about hunger today. Food Insecurity plagues a large number of people, especially in developing countries like India, Senegal, Sudan etc. The first major cause of food insecurity is farmers being dependent on rainfall, and climate change has had a huge impact on rainfall patterns. This problem has been there for a long time, but there are very few systems and areas for which the climate risk has been analyzed and preeptive measures put in motion. One such research group is the Climate Hazards Center at UC Santa Barbara. They have spent 20 years developing models to predict drought and famine for countries in Africa. Chris Funk, the director of the CHC, along with Shraddhanand Shukla has recently published a book on DEWS (Drought Early Warning Systems), called "Drought Early Warning and Forecasting". They talk about the existing DEWS in detail and it's a great guide for anyone who wants to create/improve such a system. Although they do not directly discuss Machine Learning, data is an essential part of what they do, and I am sure they are doing some kind of ML in their systems.
At Omdena, we used LSTM models to predict drought for Senegal and that made me realize how even small projects can be the start of something big. You must be thinking that predicting drought will not solve the issue. Believe me, I wondered the same for a few weeks. But then think of all the ways we could ensure that in case a drought occurs, it doesn't turn into famine. That it doesn't lead to food insecurity. There is a very fine line between drought and food insecurity, and a drought doesn't necessarily have to lead to a situation where people go to sleep hungry. And that is where policymakers come in. I honestly wish there was some business model where by companies would earn money for doing good, but this is something I have accepted that when it comes to solving wicked problems, government support is quite essential. So getting back to the point, policymakers need to direct resources to high-risk areas so that basic needs are met and people don't worry about survival. In India, more than 40% of the produces food is wasted. What if we could redirect that food to places that are experiencing high food insecurity? This actually sounds like a Math problem to me, and I am sure the engineers and data scientists of the world can solve this problem and many other similar issues.
We can empower ourselves by empowering others. And it's not some wishy-washy childhood fantasy that can never be actualized. There are people already working on these problems, and they are succeeding beautifully. It's a great time to be alive!
Happy Coding and Happy Learning!!!!