RankOOD -- Class Ranking-based Out-of-Distribution Detection
This addresses OOD detection for deep learning models, offering an incremental improvement with specific gains on a benchmark.
The paper tackles the problem of out-of-distribution (OOD) detection by proposing RankOOD, a rank-based approach that uses Plackett-Luce loss to leverage class ranking patterns, achieving state-of-the-art performance on the TinyImageNet benchmark with a 4.3% reduction in FPR95.
We propose RankOOD, a rank-based Out-of-Distribution (OOD) detection approach based on training a model with the Placket-Luce loss, which is now extensively used for preference alignment tasks in foundational models. Our approach is based on the insight that with a deep learning model trained using the Cross Entropy Loss, in-distribution (ID) class prediction induces a ranking pattern for each ID class prediction. The RankOOD framework formalizes the insight by first extracting a rank list for each class using an initial classifier and then uses another round of training with the Plackett-Luce loss, where the class rank, a fixed permutation for each class, is the predicted variable. An OOD example may get assigned with high probability to an ID example, but the probability of it respecting the ranking classification is likely to be small. RankOOD, achieves SOTA performance on the near-ODD TinyImageNet evaluation benchmark, reducing FPR95 by 4.3%.