De-Coding Computer Science

The entrepreneurial spirit is alive and coding at Truro School, thanks to Sixth Form Boarder Amari. Now in the Upper Sixth, Amari is balancing his A Levels in Maths, Computer Science, Physics and EPQ with a passion for AI data science and machine learning.

We caught up with him to learn more about his latest project, an app to help with revision.

How did you first get into computer science?

It started during Covid with the Summer Stem Institute (SSI)* that I applied to in my 4th Year. It is a six-week American programme that teaches research and data science, including AI and machine learning.

I’d done coding before in my Computer Science GCSE at Truro School, but it was quite basic. The SSI gave me the chance to develop my interest. I applied, and happily, they took me on the course with a full scholarship.

What have you done since the SSI?

The course gave me a desire to be able to create anything I wanted. I learnt the basics of AI but wanted to know about deep learning to learn how to provide a natural edge to ideas I have.

My first step was the typical image processing MNIST dataset; the digit recogniser. Being new to the self-learning process, it took me three months to fully understand, implement and deploy on docker software (used to put models into production).

Image processing led me further through my journey into learning about other data types. For example, text, tabular and audio data. Next, I plan to learn processing on video and graph data.

My initial experience with natural language processing was creating a search engine (Web Crawler). I programmed it to take a query of a STEM subject, such as “quantum physics”. The programme would allow you to search the internet for key information on that subject and output the information and potential courses that could help the person to learn more.

The web scraper required me to implement the maths behind natural language processing. I used vector mathematics to determine the similarity between words, and therefore, their relevance. I did this by calculating the angles between the vectors using cosine similarity.

Once I’d mastered that, I moved on to my app, RevisionBank.

What inspired you to create your new app?

Initially, I made it because I was too lazy to find past revision papers, so I made an exam collector tool that found them for me.

I realised that this was the start of a good idea, so I built this into a website (revisionbank.org). The application allows students to create personalised revision cards, which are then sent straight to a student’s email every 60 minutes. This helps the user to revise on the go and improve their long-term memory on a subject.

I also added a traffic light feature to enable you to rate the revision cards so that the more difficult cards are sent more often.

To complement the website, I created a chrome extension called ‘RevisionBank Snapshot’. This allows users to crop material on a webpage. The algorithm then reads the text displayed using OCR (Optical Character Recognition) and stores the information to their revision cards.

How does this fit in with your A Level choices at Truro School?

Initially, I made the website as a way to make finding revision material easier. I then realized that it would be a good project for my EPQ (Extended Project Qualification). The challenge was to create an affordable revision and tutoring system for students taking exams. I wanted to make it easier and quicker to find revision materials and make them more easily accessible on one platform. 

How has Truro School helped to support your passion for Computer Science?

Truro School has provided me with the opportunity and freedom to tinker with any crazy ideas I may have. Mr Vanstone and Mr Laity allow me to access the resources I need to create my contraptions. The school audio engineer, Mr Keary, also helped me to create the microphone from scratch for my CaesarA.I Notetaker. 

This microphone records my lessons and then translates the audio into text, before summarising it. This allows me to transcript passive notes in class so that I can listen more attentively, rather than splitting my attention between listening and taking notes.

Mr Vanstone recently gave me the Raspberry Pi Pico system to play with. I plan to integrate CaesarA.I into the Pico so that it is more portable and will transcript without a computer and send it straight to my revision cards. 

I also support other students working on their own Pico projects in Wednesday Afternoon Activities. The students are extremely keen and have made a lot of progress in a short time.

What are your hopes for the future?

I’m applying to MIT in Boston. I plan to study data science, AI recognition or cognitive science. However, I am open as to what I might find interesting during the first year.

I apply this term and am also applying to Stanford, Berkley and Harvard, so it is good to know the School is here to support me with my applications.

A huge thank you to Amari for sharing his knowledge and time with us. We feel sure that whatever the future holds for Amari, it will be an incredibly bright one.

* The Summer STEM Institute (SSI) is a six-week virtual summer program from June 19 to August 1, 2022. SSI is an international program that provides academically talented and highly motivated students the opportunity to learn through a data science and research bootcamp, a Masterclass lecture series, weekend challenges, and a mentored research project. The program was initially started by a team of graduates, students, and educators from Stanford, Harvard, MIT, and other leading universities.