Search papers, labs, and topics across Lattice.
This paper introduces Temporal Quadruple-Pattern Matching (T-QPM) to improve out-of-distribution (OOD) detection for vision-language models (VLMs) in dynamic environments. T-QPM leverages cross-modal consistency patterns between ID and OOD data by pairing OOD images with text descriptions, and learns lightweight fusion weights to adapt to temporal distribution shifts. Experiments on temporally partitioned benchmarks show that T-QPM outperforms static baselines, demonstrating robustness and temporal consistency in multimodal OOD detection.
VLMs can now better detect when they're seeing something they shouldn't, even as the world changes around them, thanks to a new method that dynamically fuses visual and textual cues.
Out-of-distribution (OOD) detection remains a critical challenge in open-world learning, where models must adapt to evolving data distributions. While recent vision-language models (VLMS) like CLIP enable multimodal OOD detection through Dual-Pattern Matching (DPM), existing methods typically suffer from two major shortcomings: (1) They rely on fixed fusion rules and assume static environments, failing under temporal drift; and (2) they lack robustness against covariate shifted inputs. In this paper, we propose a novel two-step framework to enhance OOD detection and covariate distribution shift robustness in dynamic settings. We extend the dual-pattern regime into Temporal Quadruple-Pattern Matching (T-QPM). First, by pairing OOD images with text descriptions, we introduce cross-modal consistency patterns between ID and OOD signals, refining the decision boundary through joint image-text reasoning. Second, we address temporal distribution shifts by learning lightweight fusion weights to optimally combine semantic matching and visual typicality. To ensure stability, we enforce explicit regularization based on Average Thresholded Confidence (ATC), preventing performance degradation as distributions evolve. Experiments on temporally partitioned benchmarks demonstrate that our approach significantly outperforms static baselines, offering a robust, temporally-consistent framework for multimodal OOD detection in non-stationary environments.
