Cross contrasting feature perturbation for domain generalization
Abstract Domain generalization (DG) aims to learn a robust model from source domains that
generalize well on unseen target domains. Recent studies focus on generating novel …
generalize well on unseen target domains. Recent studies focus on generating novel …
Unsupervised layer-wise score aggregation for textual ood detection
M Darrin, G Staerman, EDC Gomes… - Proceedings of the …, 2024 - ojs.aaai.org
Abstract Out-of-distribution (OOD) detection is a rapidly growing field due to new robustness
and security requirements driven by an increased number of AI-based systems. Existing …
and security requirements driven by an increased number of AI-based systems. Existing …
Resilience and security of deep neural networks against intentional and unintentional perturbations: Survey and research challenges
In order to deploy deep neural networks (DNNs) in high-stakes scenarios, it is imperative
that DNNs provide inference robust to external perturbations-both intentional and …
that DNNs provide inference robust to external perturbations-both intentional and …
Recent Advances in OOD Detection: Problems and Approaches
S Lu, Y Wang, L Sheng, A Zheng, L He… - arXiv preprint arXiv …, 2024 - arxiv.org
Out-of-distribution (OOD) detection aims to detect test samples outside the training category
space, which is an essential component in building reliable machine learning systems …
space, which is an essential component in building reliable machine learning systems …
In-or out-of-distribution detection via dual divergence estimation
Detecting out-of-distribution (OOD) samples is a problem of practical importance for a
reliable use of deep neural networks (DNNs) in production settings. The corollary to this …
reliable use of deep neural networks (DNNs) in production settings. The corollary to this …
Your Classifier Can Be Secretly a Likelihood-Based OOD Detector
J Burapacheep, Y Li - arXiv preprint arXiv:2408.04851, 2024 - arxiv.org
The ability to detect out-of-distribution (OOD) inputs is critical to guarantee the reliability of
classification models deployed in an open environment. A fundamental challenge in OOD …
classification models deployed in an open environment. A fundamental challenge in OOD …
Improving Out-of-Distribution Detection by Combining Existing Post-hoc Methods
P Novello, Y Prudent, J Dalmau, C Friedrich… - arXiv preprint arXiv …, 2024 - arxiv.org
Since the seminal paper of Hendrycks et al. arXiv: 1610.02136, Post-hoc deep Out-of-
Distribution (OOD) detection has expanded rapidly. As a result, practitioners working on …
Distribution (OOD) detection has expanded rapidly. As a result, practitioners working on …
Revisiting Score Propagation in Graph Out-of-Distribution Detection
The field of graph learning has been substantially advanced by the development of deep
learning models, in particular graph neural networks. However, one salient yet largely under …
learning models, in particular graph neural networks. However, one salient yet largely under …
Score Propagation as a Catalyst for Graph Out-of-distribution Detection: A Theoretical and Empirical Study
The field of graph learning has been substantially advanced by the development of deep
learning models, in particular graph neural networks. However, one salient yet largely under …
learning models, in particular graph neural networks. However, one salient yet largely under …