Using a corpus of 5542 fault-injected traces from 38 DL programs, the study finds a 0.19 balanced accuracy gap in fault diagnosis between within-program and cross-program evaluation caused by program-specific feature structures.
InProceedings of the 44th International Conference on Software Engineering (Pittsburgh, Pennsylvania)(ICSE ’22)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
XCheck extracts cross-layer constraints to generate test models and monitor behaviors, revealing 2,034 compiler-platform interaction bugs in three DL compilers.
An empirical study characterizing training-dynamics patterns in MLP-based software defect predictors under coupled class imbalance and overlap conditions via controlled interventions on UBD datasets.
citing papers explorer
-
Evaluation-Strategy Gap in Fault Diagnosis of Deep Learning Programs
Using a corpus of 5542 fault-injected traces from 38 DL programs, the study finds a 0.19 balanced accuracy gap in fault diagnosis between within-program and cross-program evaluation caused by program-specific feature structures.
-
Finding Compiler-Platform Interaction Bugs in Deep Learning Pipelines via Cross-Layer Constraints
XCheck extracts cross-layer constraints to generate test models and monitor behaviors, revealing 2,034 compiler-platform interaction bugs in three DL compilers.
-
Training Dynamics of Neural Software Defect Predictors under Coupled Data-Quality Issues
An empirical study characterizing training-dynamics patterns in MLP-based software defect predictors under coupled class imbalance and overlap conditions via controlled interventions on UBD datasets.