Overcoming Non-Submodularity: Towards Constant Approximation for Network Immunization
A promising theoretical breakthrough for epidemic modeling was retracted due to a critical error in its mathematical proof.
Researchers Ajitesh Srivastava and Shang-Hua Teng published a paper titled 'Overcoming Non-Submodularity: Towards Constant Approximation for Network Immunization' on arXiv. It claimed a novel method to achieve a constant-factor approximation for the complex network immunization problem, which aims to optimally select nodes to vaccinate to prevent epidemic spread. However, the authors withdrew the paper (versions v3 and v5) after discovering a fundamental flaw in the proof that invalidated the claimed approximation guarantee.
Why It Matters
Highlights the rigorous peer-review process in AI/CS research and the ongoing challenge of creating reliable algorithms for critical problems like disease containment.