DoJMA: There are many people who are apprehensive about AI. What are your thoughts about this? Is this a misconception?
Prudhvi: People worry that AI will create superhuman beings, destroy mankind and so on. I work in this industry and I see all the problems that could happen. Let us take a simple example of a self-driving car. It stops at a traffic light, all it will do is wait for the lights to cycle. But in the case of a human being, there are a lot of things that they would do, like maybe check their mobile phone, look at posters etc. The car will only do what it has been programmed to do. There are two types of AI, General and Narrow. We’ve only excelled at narrow AI, which cannot make super humans. Secondly, there is also the fear of losing jobs to automation, which is true to an extent as certain jobs will see reduction in the workforce required. However, this is not the first time that such a revolution has come around, with the Industrial Revolution being a classic example. We should embrace these changes. AI has also led to the creation of a profession known as ‘Data Labelling’, which is basically identifying the data. You cannot really predict how many jobs will be lost and how many will be created, but you cannot blame technology for this. Lastly, we come to ethics and privacy. Even I’m apprehensive of the level of intrusion into people’s privacy. For example, I texted a friend on Hangouts about a certain streaming app. A few hours later, I got an ad for buying the premium version of the same app on YouTube. I don’t have a solution for this, but I do worry about data privacy and ethics.
DoJMA: You have published many works on nature-inspired algorithms. What encouraged you to do this ?
Prudhvi: My first work that was inspired by nature was purely coincidental. I didn’t choose to do a nature inspired work. But once I did it, I got inspired to do more. I’ve realised that nature has done many more permutations and combinations than human beings, so we might as well learn from it. I also love nature; I enjoy sitting by the beach and going for long walks. Maybe that’s what inspired me at a subconscious level. It is also very interesting to think about how fish swim, how animals eat or how the universe was created; even talking about this gives me goosebumps.
DoJMA: In one of your current projects, you plan to prevent depression and its relapses. Could you share more details about this project with us?
Prudhvi: The fundamental reason I’m doing this is because AI is being deployed in a lot of places and mental health is somewhere I feel it has a lot of potential. It is also connected to my personal life. Though I have not faced any severe situations, I’ve seen a lot of people who are depressed, taking medications etc. There is also the problem of rural India having little to no access to mental healthcare, where it is still treated as a stigma. The non-availability of good psychiatrists and therapists is also a major problem. There are a lot of lifelong diseases like schizophrenia and bipolar disorder which are not being taken seriously. Although I don’t have a degree in psychology, I do read a lot on the subject, based on which I feel the way a person interacts with their phone differs depending on their mood. This can be used for effective diagnosis without the person even having to leave their room.
DoJMA: You have done multiple projects in machine learning. You must have encountered the problems of dirty data and lack of data. How do you tackle that?
Prudhvi: It’s a very interesting question. Dirty data has always been a concern when it comes to natural language processing because in this case, the data is unstructured. For example, when you download social media tweets, it is very unstructured. So, the only answer I have is patience. As far as lack of data is concerned, I have faced that a lot and the only thing that can be done is to create new data. We can create some data from trusted sources. We also have replication and creation models in machine learning, apart from web scraping, which can be done in some cases. However, if you are a passionate data scientist and your ego gets hurt that you didn’t get something, then you’ll figure out a solution to it on your own.
DoJMA: Which one of your future projects are you most excited about?
Prudhvi: Right now, I’m most excited about teaching advanced courses to a wider scale of the population. My focus is mainly on college students, startups and beginners. My aim has always been to provide affordable education to all as AI education is very expensive. I won’t spend any money on marketing and I will reduce the staff by 50 percent. This will help me in achieving my vision by reducing the costs and making it affordable for others. This is my plan for the upcoming year.