An Empirical Study on Tensor Shape Faults in Deep Learning Systems
This work addresses a critical issue for software developers using deep learning libraries, but it is incremental as it builds on existing knowledge of faults.
The study tackled the problem of tensor shape faults in deep learning systems, which are prevalent and cause crashes, by constructing SFData with 146 buggy programs and categorizing the faults into four types.
Software developers frequently adopt deep learning (DL) libraries to incorporate learning solutions into software systems. However, misuses of these libraries can cause various DL faults. Among them, tensor shape faults are most prevalent. Tensor shape faults occur when restriction conditions of operations are not met, leading to many system crashes. To support efficient detection and fixing of these faults, we conduct an empirical study to obtain a deep insight. We construct SFData, a set of 146 buggy programs with crashing tensor shape faults (i.e., those causing programs to crash). By analyzing the faults in SFData, we categorize them into four types and get some valuable observations.