Project made in 26hrs for the UCL CSSxTechSoc Hackathon (25-26/02) by Andrei Bunea, Ziyang Chen, Niladri Shekhar Dutt, and Baptiste Moalic
Education has played, and still plays a crucial role in self-development, and in establishing oneself in society. To graduate from university opens so many doors, and gives access to a living wage, stability, and growth. Nevertheless, we are not equal when facing this adventure. Not everyone has the chance to study in a calm environment, with people able to help them revise and progress. Conversely, some can pay to get an advantage through study camps and private tutors.
Needless to say, the Covid-19 crisis has had an impact even on more privileged people, that have switched off from remote classes, and that as of today have still trouble going to school.
To help democratize the access to tutors, and in a process to help struggling students to get their foot back into studying, we present: The Reviewer.
The Reviewer is a GPT-based application that adapt to the user's level and content, and provides a new way of revising by generating novel questions and assessing the user's answer to provide custom feedback.
The Reviewer can generate tailor-made questions for the user by scanning their lecture notes (either hand-written or PDFs) and using Optical Character Recognition (OCR) to create a context.
The flexibility of the GPT model then allows us to generate some questions which the user can answer. After specifying the age/level of the student, the answers by the user are then evaluated and a detailed assessment is provided on how to improve upon those answers. We can even go further and provide further tracks of reading and a learning path for the user.
Over time, the user can now integrate The Reviewer to its revision scheme and find what best works for him.
This section should list any major frameworks/libraries used to bootstrap your project. Leave any add-ons/plugins for the acknowledgements section. Here are a few examples.
In order to make the front-end run and test the program, you would need to install Node.js
Once it is done, just run:
ng serve
For testing the back-end, you can use an API tester likke Postman. For the Python program, the required modules can be found in the requirements.txt file in the back-end folder.
Once you are on the front-end web page, click "Get started". You can add your picture or pdf using the top-right "Add" button. Once it is done, please click on the centred "Start quizz". The questions will be generated, and you will have to give a complete answer to each of them.
Finally, once you've entered your answers, press "Submit", and you can take a look at the marks given by The Reviewer.
- First release
- Make an account/login system
- Add text-to-speech
- Virtual professor avatar
Our Names - Andrei Bunea, Ziyang Chen, Niladri Shekhar Dutt, Baptiste Moalic
Project Link: https://github.com/niladridutt/ucl-hackathon
Thank you to UCL Computer Science Society and Tech Society for giving us the opportunity to come up with this idea. Thank you to the industry presenters for their great ideas and help throughout the hackathon.