The paper by Joseph Goodier and Neill D. F. Campbell focuses on Out-of-Distribution (OOD) detection with Denoising Diffusion Probabilistic Models (DDPMs). OOD detection is crucial in machine learning to identify inputs significantly different from the training data, thereby avoiding incorrect predictions. The authors propose a new likelihood ratio called Complexity Corrected Likelihood Ratio, constructed using Evidence Lower-Bound evaluations from the model. This approach to OOD detection has been compared with state-of-the-art detection methods, showing comparable results.
Publication date: 26 Oct 2023
Project Page: https://arxiv.org/abs/2310.17432v1
Paper: https://arxiv.org/pdf/2310.17432
