TY - GEN
T1 - Taxonomy of Aging-related Bugs in Deep Learning Libraries
AU - Liu, Zhihao
AU - Zheng, Yang
AU - Du, Xiaoting
AU - Hu, Zheng
AU - Ding, Wenjie
AU - Miao, Yanming
AU - Zheng, Zheng
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Deep learning libraries are the cornerstone of deep learning systems, and millions of deep learning applications are built on top of deep learning libraries. Due to long-term continuous running, many numerical operations and heavy dependence on resources, deep learning libraries are prone to the effects of software aging. Aging in deep learning libraries can threaten the reliability of deep learning systems and make training and application of deep learning more time-consuming and expensive, causing users to lose confidence in it. In this work, we manually screened 138 bug reports containing aging-related bugs from a total of 13,694 bug reports in four popular deep learning libraries (i.e., TensorFlow, MXNET, PaddlePaddle and MindSpore). We analyzed the information in these 138 bug reports to answer three questions: What categories of aging-related bugs exist in deep learning libraries? What is the distribution of different categories of aging-related bugs in deep learning libraries? Which deep learning phases are most susceptible to software aging? Finally, we conducted a fine-grained taxonomy of aging-related bugs, including four levels and seventeen categories, and obtained eight important findings with corresponding practical implications.
AB - Deep learning libraries are the cornerstone of deep learning systems, and millions of deep learning applications are built on top of deep learning libraries. Due to long-term continuous running, many numerical operations and heavy dependence on resources, deep learning libraries are prone to the effects of software aging. Aging in deep learning libraries can threaten the reliability of deep learning systems and make training and application of deep learning more time-consuming and expensive, causing users to lose confidence in it. In this work, we manually screened 138 bug reports containing aging-related bugs from a total of 13,694 bug reports in four popular deep learning libraries (i.e., TensorFlow, MXNET, PaddlePaddle and MindSpore). We analyzed the information in these 138 bug reports to answer three questions: What categories of aging-related bugs exist in deep learning libraries? What is the distribution of different categories of aging-related bugs in deep learning libraries? Which deep learning phases are most susceptible to software aging? Finally, we conducted a fine-grained taxonomy of aging-related bugs, including four levels and seventeen categories, and obtained eight important findings with corresponding practical implications.
KW - aging-related bugs
KW - deep learning library
KW - deep learning phases
KW - software aging
UR - https://www.scopus.com/pages/publications/85145873502
U2 - 10.1109/ISSRE55969.2022.00048
DO - 10.1109/ISSRE55969.2022.00048
M3 - 会议稿件
AN - SCOPUS:85145873502
T3 - Proceedings - International Symposium on Software Reliability Engineering, ISSRE
SP - 423
EP - 434
BT - Proceedings - 2022 IEEE 33rd International Symposium on Software Reliability Engineering, ISSRE 2022
PB - IEEE Computer Society
T2 - 33rd IEEE International Symposium on Software Reliability Engineering, ISSRE 2022
Y2 - 31 October 2021 through 3 November 2021
ER -