A NOVEL APPROACH TO DATA RACE DETECTION USING HYBRID QUANTUM-CLASSICAL DEEP LEARNING MODELS

ABSTRACT

Concurrent programming has become fundamental in modern software systems due to the widespread use of multicore processors. However, it introduces data race conditions, which remain one of the most difficult concurrency defects to detect because of their non-deterministic behaviour and complex thread interactions. Existing static and dynamic analysis techniques often suffer from high false-positive rates, limited scalability, or significant runtime overhead. This study reviewed different existing literatures on this topic and suggests a novel (hybrid quantum-classical deep learning) approach for the early detection of data race conditions at the source-code level. It also highlights potentials of hybrid quantum-classical approach in addressing complex software reliability challenges and provides a foundation for future research in quantum-enhanced software analysis tools.

Keywords: Concurrent programming, Data race conditions, Graph Neural Networks, Quantum Machine Learning, Classical Computer, Hybrid Quantum-Classical