Speaker:Ruchira Manke

Leveraging Data Characteristics for Bug Localization in Deep Learning Programs

Deep Learning (DL) is a class of machine learning algorithms used in many applications. Like any software system, DL programs can have bugs. To support bug localization in DL programs, several tools have been proposed in the past. As most of the bugs that occur due to improper model structure known as structural bugs lead to inadequate performance during training without causing a program crash, it is challenging for developers to identify the root cause and address these bugs. To support bug detection and localization in DL programs, in this work, we propose a technique, which detects and localizes structural bugs in DL programs. Unlike the previous works, our approach considers the training dataset characteristics to automatically detect bugs in DL programs developed using two deep learning libraries, Keras and PyTorch. Since training the DL model is a time-consuming process, our approach detects these bugs at the beginning of the training process. It alerts the developer with informative messages containing the bug’s location and actionable fixes which will help them to improve the structure of the model. In our evaluation, we found that addressing these bugs lead to performance improvements in 34 out of 40 DL models.

Committee: Hridesh Rajan (major professor), Gurpur Prabhu, and Simanta Mitra