Can you classify bugs from screenshots of code?

The automatic detection and classification of bugs in code blocks committed to repositories is invaluable in software development. Traditionally, this issue has been handled through semantic analysis. Recently, a new line of research has emerged suggesting that computer vision can be used for the same task; i.e. analyzing digital images and/or screenshots of code in order to determine the presence/type of bugs! (https://dl.acm.org/doi/pdf/10.1145/3485135)

In this project you are expected to work with existing datasets and implement a variety of deep convolutional neural network architectures for the detection and classification of software bugs.