
Introducing a novel post hoc OOD detection method, NegLabel, that leverages a large set of negative labels to effectively distinguish in-distribution and out-of-distribution samples by examining their affinities towards ID and negative labels.


coremsg

leveraging-negative-labels-to-enhance-zero-shot-out-of-distribution-detection-with-vision-language-models


Leveraging Negative Labels to Enhance Zero-Shot Out-of-Distribution Detection with Vision-Language Models


title_rewrite


A unified framework that effectively leverages both visual and verbal references to improve target perception and discrimination for natural language tracking.


context-aware-multi-modal-tracking-by-natural-language-specification


Context-Aware Multi-Modal Tracking by Natural Language Specification



Leveraging emerging multimodal, vision-language foundation models as a lens to reason about and formally verify vision-based deep neural networks in terms of human-understandable concepts.


leveraging-vision-language-models-for-concept-based-analysis-and-verification-of-neural-network-classifiers


Leveraging Vision-Language Models for Concept-based Analysis and Verification of Neural Network Classifiers
