Project description: This project aims to reduce the workload of a secondary school teacher. We took a survey in Priyadarshani School to help us to understand about their workload. The survey was done in two parts:
- Teaching a class of 26 students to understand how difficult it is to teach all the students and keep them engaged simultaneously.
- We built a small questionnaire for some of our teachers about their teaching methods and assessment techniques, and some basic questions about what they would want from a teaching assistant.
This teaching assistant's main goal is to facilitate the teaching process by helping teachers obtain thorough lesson materials and notes. It also provides an advanced system for assessing student replies to open-ended questions by using natural language processing (NLP) algorithms to examine and contrast the answers with those supplied by the teacher. This review technique is significant because it minimizes bias and places a high priority on basic factual correctness, guaranteeing impartial and fair assessment procedures.
We have created a system that takes some basic inputs to create a lesson plan to teach in class.
We fine-tuned the Mistral 7B instruct model and added CBSE textbooks as a knowledge database so it can be used as a reference.
To create this automated grading and reviewing system we first score the sheets and then give reasons for the marking (for every mark deducted), to check the answer sheets first requires a modal answer sheet as a reference to scoring another sheet, we want to ensure to understand the perspective of the student and improve them at every step they need to, as every teacher cannot create a detailed analysis of every student, So this model scores the reviews and remarks in the database which helps to create scopes of improvement for every student.