Automatic student performance prediction makes AI-based teaching much more advanced than traditional teaching. Fundamental hands-on skills and key algorithms will be introduced and summarized for the problem of students’ learning level regression. Special focus will be cast on the deep learning method. This workshop provides a wide range of possible applications for different disciplines, like students’ performance prediction, automatic warning/intervention systems, and individualized education systems. This workshop is developed specifically for the participants without any technology or programming background.

​

• In this workshop, participants will be familiar with the basic programming environment setup and user interface of the free computing platform COLAB.

• Basic algorithms in regression will be introduced, including Linear Regression, Random Forest, Support Vector Machine, and Deep Learning.

• The workshop will also lead the participants to build an Artificial Neural Network (ANN) model in a toy example.

• The concepts of pre-trained models and fine-tuning will be introduced. Participants will learn to fine-tune a pre-trained model with custom datasets.

• Participants are expected to understand which factors could affect the model performance.