Deep Knowledge Tracing with AWS Neuron

This repository implements Deep Knowledge Tracing (DKT) using PyTorch and AWS Inferentia2 for efficient inference on SageMaker.

Overview

Deep Knowledge Tracing is a machine learning method that models student knowledge over time. This implementation uses AWS Neuron SDK for optimized inference on AWS Inferentia2 accelerators.

In this repository, ASSISTment2009 "skill-builder" dataset are used. You need to download the dataset on the following path:

datasets/ASSIST2009/

Features

• Implementation of DKT model using PyTorch
• AWS Neuron optimization for Inferentia2
• Support for ASSIST2009 dataset
• Real-time inference monitoring

Requirements

• Python 3.8+
• PyTorch 1.12+
• torch-neuronx
• numpy
• pandas
• scikit-learn

Installation

# Clone the repository
git clone https://github.com/yourusername/dkt-neuronx.git
cd dkt-neuronx

Usage

run dkt_model.ipynb or dkt_model_ko.ipynb(Korean) on SageMaker Notebook instance (inf2 or trn1).

Model Architecture

The DKT model uses LSTM to trace student knowledge states:

• Input embedding layer
• LSTM layer
• Output layer with sigmoid activation

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.