Adaptive Prompt Learning via Gaussian Outlier Synthesis for Out-Of-Distribution Detection

Out-of-distribution (OOD) detection aims to distinguish whether detected objects belong to known categories or not. Existing methods extract OOD samples from Indistribution (ID) data to regularize the model's decision boundaries. However, the decision boundaries are not adequately regularized because the model does not have sufficient knowledge about the distribution of OOD data. To address the above issue, we propose an Adaptive Prompt Learning framework via Gaussian Outlier Synthesis (APL-GOS) for OOD detection. Specifically, we leverage the Vision-Language Model (VLM) to initialize learnable ID prompts by sampling standardized results from pre-defined Q&A pairs. Region-level prompts are synthesised in lowlikelihood regions of class-conditional gaussian distributions. These prompts are then utilized to initialize learnable OOD prompts and optimized with adaptive prompt learning. Also, OOD pseudo-samples are synthesised via gaussian outlier synthesis. The aforementioned methodology regularizes the model to learn more compact decision boundaries for ID and OOD categories. Extensive experiments show that APLGOS achieves state-of-the-art performance with less ID data on four mainstream datasets.