Advances in Probabilistic Systems and Software Reliability

The field of probabilistic systems and software reliability is moving towards the development of innovative methods for testing, debugging, and validating complex systems. Researchers are exploring new approaches to differentiate between software bugs and hardware noise, amplify the occurrence of elusive bugs, and detect physics failures in simulation-based software. A key direction is the integration of probabilistic and physical models to improve the fidelity of simulations and enable more realistic interactions between systems and their environments. Noteworthy papers include: A Model-Independent Theory of Probabilistic Testing, which proposes a general framework for probabilistic testing, and Distinguishing Quantum Software Bugs from Hardware Noise, which introduces a statistical approach to differentiate between quantum software bugs and hardware noise. Black-Box Bug-Amplification for Multithreaded Software presents a novel approach to amplify the occurrence of concurrency bugs, while Runtime Failure Hunting for Physics Engine Based Software Systems provides a large-scale empirical study on physics failures in simulation-based software. Axioms for Model Fidelity Evaluation establishes fundamental axioms for evaluating model fidelity, and Half-Physics enables kinematic 3D human models to physically interact with their environment.

Sources

A Model-Independent Theory of Probabilistic Testing

Distinguishing Quantum Software Bugs from Hardware Noise: A Statistical Approach

Black-Box Bug-Amplification for Multithreaded Software

Runtime Failure Hunting for Physics Engine Based Software Systems: How Far Can We Go?

Axioms for Model Fidelity Evaluation

Half-Physics: Enabling Kinematic 3D Human Model with Physical Interactions

Built with on top of