Taming Score-Based Denoisers in ADMM: A Convergent Plug-and-Play Framework

1. 1Identify the challenges of integrating score-based generative models into ADMM
2. 2Develop a plug-and-play framework to address the mismatch between noisy data manifolds and ADMM iterates
3. 3Establish convergence guarantees for the proposed framework
4. 4Implement and test the framework on various inverse problems

Machine learning researchers and engineers working on inverse problems and optimization algorithms can benefit from this framework, as it provides a convergent and efficient way to integrate score-based denoisers into ADMM

💡 The proposed framework addresses the challenges of integrating score-based generative models into ADMM, providing a convergent and efficient solution for inverse problems

💡 Convergent plug-and-play framework for integrating score-based denoisers into ADMM