The authors aim to explore the prevalence, introduction, removal, and survivability of ML-specific code smells in ML-enabled systems through a large-scale empirical study.
