From creating a moonshot solution to my Neurosymbolic AI project
An honest update about the last two months
Overview of the Month 💯
Moonshot Challenge (UltraRice)
The three parts of a moonshot idea are: A huge problem. A radical solution. A breakthrough technology.
My team, Anika Gandhi, Ari Dutilh, Laasya Chevendra, Orsi Benkő, and I worked on creating a solution having all these parts for our very own moonshot idea as part of our TKS journey. Our presenter Anika presented the project and we won the "Make It Real" and People's Choice awards.
Our moonshot idea aims to revolutionize the nutritional value of rice itself by focusing on enhancing its inherent properties. Through the implementation of ultrasonic treatment, we can initiate a series of controlled reactions within the rice grains. This precise and non-invasive process unlocks the potential to augment vital nutrients, vitamins, and minerals. By bridging the gap between production and nourishment, our solution empowers individuals with more nutrient-dense rice to combat malnutrition effectively.
Throughout the journey of coming up with the solution, the key lesson I learned was "perspective". While listening to Astro Tellers' talks, one of the key things I understood from his talk was that "Shifting your perspective is better than being smart". He iterated on how finding a different perspective would lead to more innovative and creative solutions.
We are currently looking into building this as a company and extensively looking for resources to do this, so any help would be appreciated!
Neurosymbolic AI project at the TKS Showcase
Last month, I created an one pager and video to be displayed at the TKS Showcase (An event for everyone to view the work done at TKS).
Video for Showcase:
A road trip away from Chennai
For people across the world reading this, the temperatures in Chennai, India are the north of 35 C every day, so going to a hill station, where the temperature was an average of 18 C, was a refreshing change of pace and a pretty good way to beat the heat.
Music
One thing I wanted to do before going to university in August was to write a song, while I kinda have half of one song ready (I got a verse and a chorus for it), I decided to try out some new tracks and here is a combination of some of my favourites so far.
The lesson of the month 🌟
To be honest, I have had a bunch of realizations these last few months, and one of the most important one is this: the future is inherently uncertain, and that uncertainty can be quite daunting and frightening (I promise I don’t mean this in like a very bad way, Im ramping up to something better). However, it is really important to focus on the present moment and understand that we don’t actually have total control over the future. External factors and unexpected events can influence our lives. Instead, I believe we should make choices and take action in the present to shape our lives and create a better future. We have gotta prepare for the future, by setting up a proper plan. We can't control everything, but what we can control is our responses and attitudes. By being proactive, resilient, and optimistic, we navigate life's changes with confidence.
AI/ML Project Updates (My TKS Focus) ⚙️ ⚽️
As I mentioned in the previous issues, I started my focus on Artificial Intelligence 🧠 at TKS. I am passionate about the application of Artificial intelligence (AI) ⚙️ in the real world (particularly in the football industry). As part of the process, I have written my learn article on the impact of AI ⚙️ in the football industry titled “How Artificial Intelligence(AI) could improve the way $4.4 billion 💵 is spent on football?” on Medium.
For my replicate 1, instead of creating a distributed deep-learning model for movie recommendations, I opted for a more approachable solution. I built a hybrid neural network for movie recommendations and shared my journey in a detailed article on Medium. In the article, I explore how a hybrid neural network enhances personalized suggestions by combining collaborative filtering and deep learning techniques. I discuss the process of building the network using TensorFlow Recommenders and demonstrate accurate movie recommendations through training.
This time, for my replicate 2, I created a neuro-symbolic AI system that generate mock-tails for users based on a users preferred flavour profiles. While the project technically doesn’t have much of a real world impact, it does showcase the potential of neurosymbolic AI in the real world. This project uses a simple rule based if-then system to remove all infeasible combinations, and then a neural network to predict which mocktail combinations would be the best for the users request. I wrote an article about the entire process and explained the everything in detail in my video as well
Video link:
Something interesting this month 💫
As the football transfer season is starting and there are many big money moves starting up, I am going to spend some time looking into some of the young standout performers from the Big 5 leagues, about their transfer status and their future. I am also going to explore the current transfers that are being made, and look deeper into the future of their transfer. So follow my Twitter for updates on that: https://twitter.com/PrabhatM27
Here is a sneak peak about what to expect:
Conclusion 🔴
Thank you for reading all this way, and I am trying to be consistent in these updates and i look forward to sending the next update soon!
If you haven’t already, make sure to check out my latest article about creating a Disease Prediction symptom from symptoms using Neuorsymbolic AI! 📄