The field of OOD research encompasses several key areas: 🔍 OOD detection aims to identify when a model is presented with inputs that deviate from its training distribution. This allows systems to flag unusual cases for human review or fallback strategies.

Out of Distribution (OOD) Detection is the task of detecting instances that do not belong to the distribution the classifier has been trained on. OOD data is often referred to as "unseen" data, as the model has not encountered it during training.

2021年10月21日 · In this survey, we first present a unified framework called generalized OOD detection, which encompasses the five aforementioned problems, i.e., AD, ND, OSR, OOD detection, and OD. Under our framework, these five problems can be seen as special cases or sub-tasks, and are easier to distinguish.

In this survey, we first present a unified framework called generalized OOD detection, which encompasses the five aforemen-tioned problems, i.e., AD, ND, OSR, OOD detection, and OD. Under our framework, these five problems can. be seen as special cases or sub-tasks, and are easier to distinguish.

A Python library for Out-of-Distribution (OOD) Detection with Deep Neural Networks based on PyTorch. The library provides: Out-of-Distribution Detection Methods; Loss Functions; Datasets; Neural Network Architectures, as well as pre-trained weights; Data Augmentations; Useful Utilities

2024年11月11日 · Out-of-Distribution (OOD) detection refers to identifying data that differs significantly from the distribution on which a machine learning model was trained, known as the in-distribution (ID). To understand OOD detection, let’s break down the concept: What is Out-of-Distribution Detection? Why is Out-of-Distribution Detection Important?

2024年7月31日 · In this survey, we first present a generalized OOD detection v2, encapsulating the evolution of AD, ND, OSR, OOD detection, and OD in the VLM era. Our framework reveals that, with some field inactivity and integration, the demanding challenges have become OOD detection and AD.

A new report which provides benchmarking results on ImageNet and for full-spectrum detection. A unified, easy-to-use evaluator that allows evaluation by simply creating an evaluator instance and calling its functions.

We address the problem of out-of-distribution (OOD) detection for the task of object detection. We show that residual convolutional layers with batch normalisation pro-duce Sensitivity-Aware FEatures (SAFE) that are consis-tently powerful for distinguishing in-distribution from out-of-distribution detections.

Motivated by this observation, we propose a novel OOD scoring method named Virtual-logit Matching (ViM), which combines the class-agnostic score from feature space and the In-Distribution (ID) class-dependent logits.