Automated Validation of Large-scale Software Systems (NUS)


More information here

This talk presents an overview of our research on automated software validation (SV), especially methods/tools to support developers during software testing and debugging activities. The talk presents our application of input grammars for testing (a) the correctness of traditional software systems (e.g., compilers/interpreters like Mozilla Rhino and Google’s Closure) and (b) the fairness properties of Machine Learning based systems (e.g., NLP systems like Google BERT, StanfordNLP and AllenNLP). Finally, the talk will conclude with an outline of our ongoing work on building intelligent (or data-centric) SV methods, human-in-the-loop SV methods, and SV methods for interdisciplinary software systems.