My Podcasting Experience (Part IV): Industry Experts
“The beautiful thing about podcasting is that it’s just talking. It almost has no definitive form and is one of the best ways to explore an idea.” — Joseph Rogan, comedian and podcast host
Welcome, friends, to part 4 of my ACM/IRISS podcast article series. By now, you must have got the hang of the flow of this season. Each write-up is unique yet connected with all others based on the three significant computer science events that happened at IIT Gandhinagar recently — IRISS, ACM-W workshop, and ACM-annual event. This part contains a few interestingly educative sneak peeks from my podcasting sessions with some of the leading experts in the industry.
So, without further ado, here come the snippets!
Nikhil Rasiwasia: Applied Science Manager at Amazon
In your opinion, how can we promote computing education and research in the present times?
I think that the process of making anything interesting for an individual should start from a very early age. To develop that passion, we should design activities that focus on questions like how to make things fun? (Similarly, in terms of CS) How to make coding fun? For example, a friend of mine started a company for kindergartens in Mumbai, and it is all about how the 3–5 years old kids can come and understand what coding is, in a very simple way! Through this, they would eventually grasp what they can do with coding. There is a dire need to change the ‘coding/ computing is meant only for boys’ kind of stereotypes. Slowly, people are starting to overcome this gender bias, but we still need to work on it actively. If there is a requirement to educate parents on that front, then that should be the second initiative to make computer science more pervasive in our lives.
Machine Learning — one of the biggest buzz words in the field of computer science today. How do you perceive its future?
I think anybody who has tried to predict machine learning has failed, so let me not do that job (chuckles!). But, I do see it growing with a lot of new paradigms coming in. As of now, Supervised Learning is what took the most significant piece of mindshare for the last decade. At some level, people are itching for more. From a research point of view, the question is, how can we leverage a lot of data which is not labeled? Currently, ML has been very effective with the labeled data. In the last 2–3 years, people have also found inventive ways of utilizing unlabeled data. A few days ago, Microsoft released something called the T-NLG, that’s Turing Natural Language Generation, which can generate almost human-looking sentences and paragraphs from thin air. In essence, it looks like this model understands what language is, without any supervision. So, these are large scale efforts in understanding how the world works, and I think this is going to take the mainstream in the next couple of years.
AI is another burning topic in the area of computing now-a-days. In terms of spreading awareness among people from other disciplines or the laypeople, how would you explain the connection or relation of machine learning with artificial intelligence?
It’s a good question. There is no universal definition, but to me, artificial intelligence is that broad-spectrum which encompasses all intelligent systems. If this system is handcrafted (based on rules), it is known as rule-based AI. The other part of AI is when you create architectures for machines to infer these rules on their own. So, ML is a machine learning a set of rules or patterns. Therefore, AI is a superset of machine learning. Now, within ML, a particular kind of architecture to learn these rules is called neural networks. If these architectures have multiple layers or the depth of these architectures is vast, that’s what deep learning is all about! To conclude, AI is a superset of ML, which in turn, is a superset of neural networks, which again is a superset of deep learning.
R Venkateswaran: Holder of different key positions at Persistent Systems and Speaker of ACM-India
How would you describe your work to the general public?
I head the industrial and IoT business at Persistent. It is the application of technology, primarily digital technology, to transform the specific industries, mainly manufacturing. Today, manufacturing industries are very labor-intensive, there are a lot of areas where inefficiencies creep in, and errors or quality issues happen. The adoption of technology can address all these points, and that’s what we do as a company. We try to provide solutions for some of these challenges.
You were a part of an engaging panel conversation during IRISS 2020 at IIT Gandhinagar. Can you please inform our listeners about some crucial points discussed during this session?
Focused on how do we build and nurture a research organization, this panel had members from different industries who are well-known for their contributions to research. All these organizations work very closely with academia. We discussed how some of the areas that industries work on are also the practical problems that we can bring to the research community, because, what we observe sometimes is that research is one level of abstraction away from the actual issues. So, working closely with industry will help to bridge that gap. We also talked about how to make sure that faculty working in current research areas can contribute to the problems of the industry. Along the same lines, how industrial research can bring in a lot more context to the research domains of academia.
In short, the industry should be the practical application of academic research.
Yogesh Kulkarni: Over 20 years of experience in Engineering Software Development and Principal Architect in CTO Office of Icertis
What is your take on artificial intelligence, one of the most trending topics in the present world?
Artificial Intelligence has really solved many core problems that we are facing. Particularly in the field of image processing and slowly in the domain of natural language processing. Imagine, doing translations between languages — it was very, very crude previously, but now, machine-learning-based machine translations are coming pretty close to what humans do. However, the wings of AI are enormous, and a lot still remains to be discovered, explored, and deciphered.
What is your advice for students?
I am going to give what is probably a slightly philosophical answer. I think you have to find what is known as Ikigai — your own life’s purpose. Pick an area, a problem which you like, and you should be good at it. That’s the first thing. The next is that this problem should align with the need of the world — it should be a requirement for society. The last crucial thing is that you should get paid for it. If a domain or interest satisfies all the criteria mentioned above, students should definitely go for it.
So, time to say bye-bye now. I will bring some more engaging tit-bits, from the remaining detailed chit-chat versions, for you soon. A little teaser — in the upcoming article, you will get a glimpse into the world of some eminent researchers in computing! The podcasts will also be available after some time. Till then, happy computing to all.
(For Part 3 of this series, please visit here. Part 5 can be found here.)
