Using The Power Of AI For Good
— AI, IoT, Deep Learning — 4 min read
Meet Shikhar Kwatra
Artificial Intelligence-IoT IP Engineer @ IBM
Durham, North Carolina
An Inventor of 200+ Patent Applications.
Continuing to work on emerging areas and work on innovative ideas. He is currently speaking on various occasions to inspire fellow engineers to think from an inventive mindset and develop new products.
His major research interests include AI, the Internet of Things, Computer Architecture and Embedded Systems and the study of behavioral patterns.
Shikhar projects on Artificial Intelligence can be found on his YouTube channel.
How you describe your first encounter with Ai?
While doing my master’s in electrical engineering, I was more focused on applying my knowledge of hardware and integrating with different pieces of software to create different applications and was hearing about the potential of AI from some folks specializing in the field of Machine Learning.
After I joined IBM, I started working on different kinds of certifications related to machine learning and deep learning and realized the power of using AI to teach machines to become smarter.
That curiosity and drive of embedding intelligence into machine got me excited to dive into the field and create several proofs-of-concept (PoCs) using the infusion of AI and IoT.
How did you become an AI developer (or AI architect position)?
After completing a ton of certifications in the field of AI, I started applying those skills and creating practical applications, some of which were started as part of my project Make-Happy initiative.
They were demonstrated in AI-Hackathon, KidsforCode initiative and Marbles museum to showcase the power of Applied AI in areas pertaining to Computer vision and Natural Language Processing.
Gaining a wonderful experience and insight from those initiatives instilled the confidence in me to take the leap into applying AI across different domains and officially make the transition to Data and AI architect position.
How do you define Intelligence?
Intelligence is a very subjective term honestly and some people can get carried away talking about it (just like me).
Intelligence, to me, can be considered as an amalgamation of curiosity and passion with a focus to dive into a certain topic, be it their field of study or taking care of a logical matter at hand.
It doesn’t necessarily need to be scoped down to the field of AI.
The ability of a human or a machine to think about a problem from different perspectives or arrive at a solution by broadening their horizon of thinking can certainly be a sign of intelligence.
I think Intelligence can also be correlated with the ability to grasp information and make use of said digested or processed information.
The only thing I caution people about the interpretation of intelligence is not to put a halt to their thought process or think simply in a unidirectional manner, which can be a sign of dogmatism.
As you can see, I clearly digressed but I believe the day you would stop learning or processing new perspectives or even different pieces of information would be the day you might cease to think or live (and merely simply exist).
Can you mention any interesting AI projects you have been involved in?
As part of my make-happy project brand, I developed a mood classifier enabled custom chat notifier system.
Basically, the system starts with monitoring user’s facial expressions (via web-camera or can potentially be any IP camera connected to a hardware processor or assist processor capable of running ML or Deep learning algorithms) and computes the emotion probability variations over time.
Once the emotion probability is detected beyond a certain threshold, the client-server architecture polls the messaging API to send a custom-tailored notification to the user while continuing to monitor said user’s mood variations.
If a different emotion is detected, a modified notification is sent to the user’s mobile device to alter or amplify the suspected mood based on learning said emotional variations through the Haar cascade classifier.
I was lucky to have got a patent out of this invention as well!
Real-time Mood analysis-based Notifier mechanism- A custom-tailored notification is received by the user based on varying moods.
What advice would you give to a student in the field of artificial intelligence?
I would suggest that doing certifications in the field of AI is one thing and applying those skills from a practical standpoint is another.
Always learn to apply your skills outside of your comfort zone, that is, not just focus on solving problems present in your course assignments/coursework but also look at practical problems, even if that might require some extra legwork of pulling unstructured data or mining additional data for building the ecosystem.
For instance, if you worked on a project as part of coursework wherein partial information was already provided, say, some steps of the dataset or data which has already been scraped and pre-processed, that will surely teach you how to deal with refined problems but I’d encourage you to search for problems which can be created from scratch.
You don’t have to reinvent the wheel but once you learn all the steps involved in the pipeline, you’ll be able to better understand the nuances and techniques involved in solving the problem.
Also, as indicated above, working on Kaggle problems is certainly good for practicing but eventually, from a practical standpoint, in order to attain an even more refined understanding of the subject matter, it would be better to tackle problems from a specificity standpoint.
i.e., If there is a field of AI you’re interested in, say text analytics, speech processing/analysis or even computer vision, try to expand your skills in that area and then you can move to other focus areas perhaps.
Note: You can make more descriptive information which can be included in your website:
For more information about me from Inventing standpoint:
IBMer Stories : 3 Personality Traits Observed in Prolific Inventors
My Inventions @ United States Patent and Trademark Office (USPTO).
I launched this new initiative to encourage people to enter AI and in the field of STEM:
Please take a look at my implemented ideas below where I make use of AI to come up with novel solutions:
- Audio summarizer system where system condenses a huge document link (patent for eg.) into 10-20 lines and then integrated with Watson Text to Speech to articulate the summary.
- Dynamic GIF creation using 3D projections!
To learn more about Shikhar Kwatra AI-IoT infusion conferences and projects can be contacted through his website