Detecting TensorFlow Program Bugs in Real-World Industrial Environment
Deep learning has been widely adopted in industry and has achieved great success in a wide range of application areas. Bugs in deep learning programs can cause catastrophic failures, in addition to a serious waste of resources and time.
This paper aims at detecting industrial TensorFlow program bugs. We report an extensive empirical study on 12,289 failed TensorFlow jobs, showing that existing static tools can effectively detect 72.55% of the top three types of Python bugs in industrial TensorFlow programs. In addition, we propose (for the first time) a constraint-based approach for detecting TensorFlow shape-related errors (one of the most common TensorFlow-specific bugs), together with an associated tool, \textsc{ShapeTracer}. Our evaluation on a set of 60 industrial TensorFlow programs shows that \textsc{ShapeTracer} is efficient and effective: it analyzes each program in at most 3 seconds and detects effectively 40 out of 60 industrial TensorFlow program bugs, with no false positives. \textsc{ShapeTracer} has been deployed in the \AnonP platform (\emph{anonymized for double blind review}) and will be released soon.
Tue 16 NovDisplayed time zone: Hobart change
12:00 - 13:00 | ProgrammingJournal-first Papers / Research Papers / NIER track at Kangaroo Chair(s): Amiangshu Bosu Wayne State University | ||
12:00 20mTalk | Detecting TensorFlow Program Bugs in Real-World Industrial Environment Research Papers Chen Liu , Jie Lu SKL Computer Architecture, ICT, CAS, Guangwei Li Institute of Computing Technology, Ting Yuan SKL Computer Architecture, ICT, CAS University of Chinese Academy of Sciences, China, Lian Li Institute of Computing Technology at Chinese Academy of Sciences, China, Feng Tan Alibaba Group, Jun Yang Alibaba Group, Liang You Alibaba Group, Jingling Xue UNSW Sydney Pre-print | ||
12:20 20mTalk | Why Do Developers Remove Lambda Expressions in Java? Research Papers Mingwei Zheng Huazhong University of Science and Technology, Jun Yang Huazhong University of Science and Technology, Ming Wen Huazhong University of Science and Technology, Hengcheng Zhu The Hong Kong University of Science and Technology, Yepang Liu Southern University of Science and Technology, Hai Jin Huazhong University of Science and Technology | ||
12:40 10mTalk | A Mocktail of Source Code Representations NIER track Dheeraj Vagavolu RISHA Lab, Indian Institute of Technology, Tirupati, Karthik Chandra Swarna RISHA Lab, Indian Institute of Technology Tirupati, Sridhar Chimalakonda RISHA Lab, Indian Institute of Technology, Tirupati | ||
12:50 10mTalk | On Tracking Java Methods with Git Mechanisms Journal-first Papers Yoshiki Higo Osaka University, Shinpei Hayashi Tokyo Institute of Technology, Shinji Kusumoto Osaka University | ||