Taxonomy of real faults in deep learning systems
Nargiz Humbatova, Gunel Jahangirova, Gabriele Bavota, Vincenzo Riccio, Andrea Stocco, Paolo Tonella
Abstract
The growing application of deep neural networks in safety-critical domains makes the analysis of faults that occur in such systems of enormous importance. In this paper we introduce a large taxonomy of faults in deep learning (DL) systems. We have manually analysed 1059 artefacts gathered from GitHub commits and issues of projects that use the most popular DL frameworks (TensorFlow, Keras and PyTorch) and from related Stack Overflow posts. Structured interviews with 20 researchers and practitioners describing the problems they have encountered in their experience have enriched our taxonomy with a variety of additional faults that did not emerge from the other two sources. Our final taxonomy was validated with a survey involving an additional set of 21 developers, confirming that almost all fault categories (13/15) were experienced by at least 50% of the survey participants.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
55 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
Cited by top-tier papers55
- Generative Models for Effective ML on Private, Decentralized DatasetsSean Augenstein, H. Brendan McMahan, Daniel Ramage, Swaroop Ramaswamy et al.ICLR 2020 · 207 citations
- Model-based exploration of the frontier of behaviours for deep learning system testingVincenzo Riccio, Paolo TonellaFSE 2020 · 134 citations
- Collaboration Challenges in Building ML-Enabled Systems: Communication, Documentation, Engineering, and ProcessNadia Nahar, Shurui Zhou, Grace A. Lewis, Christian KästnerICSE 2022 · 122 citations
- A comprehensive study on challenges in deploying deep learning based softwareZhenpeng Chen, Yanbin Cao, Yuanqiang Liu, Haoyu Wang et al.FSE 2020 · 121 citations
- DeepCrime: mutation testing of deep learning systems based on real faultsNargiz Humbatova, Gunel Jahangirova, Paolo TonellaISSTA 2021 · 114 citations
Related papers
- An Empirical Study on Deployment Faults of Deep Learning Based Mobile ApplicationsZhenpeng Chen, Huihan Yao, Yiling Lou, Yanbin Cao et al.ICSE 2021 · 73 citations
- Towards Understanding the Faults of JavaScript-Based Deep Learning SystemsLili Quan, Qianyu Guo, Xiaofei Xie, Sen Chen et al.ASE 2022 · 13 citations
- Repairing deep neural networks: fix patterns and challengesMd Johirul Islam, Rangeet Pan, Giang Nguyen, Hridesh RajanICSE 2020 · 102 citations
- Understanding performance problems in deep learning systemsJunming Cao, Bihuan Chen, Chao Sun, Longjie Hu et al.FSE 2022 · 33 citations
- Why Attention Fails: A Taxonomy of Faults in Attention-Based Neural NetworksSigma Jahan, Saurabhsingh Rajput, Tushar Sharma, Masud RahmanICSE 2026
