Ood Detection

Out of Distribution (OOD) Detection refers to the task of identifying data instances that do not belong to the distribution the classifier was trained on. OOD data is often referred to as "unseen" data because the model has not encountered it during training. The goal of this task is to train a model to distinguish between in-distribution (ID) data, which the model has seen during training, and out-of-distribution (OOD) data, which it has not encountered, in order to enhance the model's robustness and generalization capabilities. This can be achieved by training an independent OOD detector or by modifying the model architecture and loss function. In the field of computer vision, OOD detection is particularly valuable for identifying anomalies and unknown objects.