Automated Essay Scoring System

Automated Essay Scoring System (CNN + LSTM + Attention)

Developed an AI-powered system to automatically grade essays based on factors such as coherence, vocabulary richness, grammar, and topic relevance. The solution was trained using the Kaggle ASAP Automated Essay Scoring Dataset, applying advanced NLP and deep learning techniques.

The system combines:

LSTM (Long Short-Term Memory) networks for capturing sequential and contextual patterns in text

CNN (Convolutional Neural Networks) for extracting local linguistic features

Attention mechanisms for focusing on the most relevant parts of each essay

Key Features:

Automated grading with high consistency and reduced human bias

Text preprocessing pipeline: tokenization, stopword removal, normalization

Feature extraction using n-grams and word embeddings

Supports fast scoring for large volumes of essays

Scalable for integration into educational platforms or assessment tools

Impact:

This system streamlines the evaluation process for educators, improves feedback speed for students, and can be adapted for different languages and writing standards.