An Analysis of the Relationship between Conditional Entropy and Failed Error Propagation in Software Testing
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 13920
- Type
- E - Conference contribution
- DOI
-
10.1145/2568225.2568314
- Title of conference / published proceedings
- Proceedings. 36th International Conference on Software Engineering, 2014. ICSE 2014.
- First page
- Accepted
- Volume
- -
- Issue
- CONFCODENUMBER
- ISSN
- 0270-5257
- Open access status
- Out of scope for open access requirements
- Month of publication
- March
- Year of publication
- 2014
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
-
- Citation count
- 23
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper is the first to demonstrate how information theoretic models can minimise false negatives for oracle decisions during software testing. We demonstrated the predictive power of our model on real numerical programs with seeded faults. We witnessed 0.95 Spearman correlation between entropy loss measures and the probability of failed error propagation across 30 programs using 7.5 million random test cases. The experiment methodology was the foundation for EPSRC projects EP/P005888/1 and EP/P006116/1, leading to further papers applying information theoretic models. The Failed Error Propagation model has since been extended to program robustness, winning funding from FaceBook.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -