Improving Out-of-Distribution Detection with Markov Logic Networks
This work addresses reliability issues for deep learning models in open-world scenarios, offering an incremental improvement with added explainability.
The paper tackled the problem of out-of-distribution detection in deep learning by augmenting existing detectors with Markov logic networks, resulting in significant performance enhancements across multiple datasets while maintaining computational efficiency.
Out-of-distribution (OOD) detection is essential for ensuring the reliability of deep learning models operating in open-world scenarios. Current OOD detectors mainly rely on statistical models to identify unusual patterns in the latent representations of a deep neural network. This work proposes to augment existing OOD detectors with probabilistic reasoning, utilizing Markov logic networks (MLNs). MLNs connect first-order logic with probabilistic reasoning to assign probabilities to inputs based on weighted logical constraints defined over human-understandable concepts, which offers improved explainability. Through extensive experiments on multiple datasets, we demonstrate that MLNs can significantly enhance the performance of a wide range of existing OOD detectors while maintaining computational efficiency. Furthermore, we introduce a simple algorithm for learning logical constraints for OOD detection from a dataset and showcase its effectiveness.