Grading Write-Ups Using Deep Learning

EasyChair Preprint 15704

Abstract

In most of the countries, the quality of education in the urban areas and the rural ar-eas have huge gaps. As a result, the evaluation of a piece of text, specifically hand-written text is very subjective and at times can be biased. So, this project aims to build an unbiased grading system using Deep Learning and Transfer Learning. The aim of the paper is to show - how Transfer Learning in the form of pre-trained word embed-dings can be introduced with Deep Learning, a regression model to predict a score ac-curately on a scale of 1 – 10 and to compare the performance of the model on real world data. This paper shows an approach to Automated Essay Grading using Deep Learning and Transfer Learning. The aim of the paper is to show how Transfer Learn-ing in the form of pre-trained word embeddings can be introduced with Deep Learn-ing; A regression model to predict a score accurately on a scale of 1 – 10; and to compare the performance of the model on real world data.

Keyphrases: Automatic Essay Grading, Project Essay Grade, ReLU, Transfer Learning, data cleaning, deep learning, exploratory data analysis, feature extraction

@booklet{EasyChair:15704,
  author    = {Siladitya Mukherjee and Avijit Chakraborty and Anal Acharya},
  title     = {Grading Write-Ups Using Deep Learning},
  howpublished = {EasyChair Preprint 15704},
  year      = {EasyChair, 2025}}