A survey of mix-based data augmentation: Taxonomy, methods, applications, and explainability

C Cao, F Zhou, Y Dai, J Wang, K Zhang - ACM Computing Surveys, 2024 - dl.acm.org
Data augmentation (DA) is indispensable in modern machine learning and deep neural
networks. The basic idea of DA is to construct new training data to improve the model's …

Endoood: Uncertainty-aware out-of-distribution detection in capsule endoscopy diagnosis

Q Tan, L Bai, G Wang, M Islam, H Ren - ar** functions
Q Bouniot, P Mozharovskyi, F d'Alché-Buc - arxiv preprint arxiv …, 2023 - arxiv.org
Data augmentation is an essential building block for learning efficient deep learning models.
Among all augmentation techniques proposed so far, linear interpolation of training data …

Being confident in confidence scores: calibration in deep learning models for camera trap image sequences

G Dussert, S Chamaillé‐Jammes… - Remote Sensing in …, 2023 - Wiley Online Library
In ecological studies, machine learning models are increasingly being used for the
automatic processing of camera trap images. Although this automation facilitates and …

On Uncertainty In Natural Language Processing

D Ulmer - arxiv preprint arxiv:2410.03446, 2024 - arxiv.org
The last decade in deep learning has brought on increasingly capable systems that are
deployed on a wide variety of applications. In natural language processing, the field has …

Revisiting Interpolation Augmentation for Speech-to-Text Generation

C Xu, J Wang, X Liu, Q Dong, C Zhang, T **ao… - arxiv preprint arxiv …, 2024 - arxiv.org
Speech-to-text (S2T) generation systems frequently face challenges in low-resource
scenarios, primarily due to the lack of extensive labeled datasets. One emerging solution is …

Uncertainty Quantification Based on Gaussian Processes for Image Segmentation Tasks

B Gao, R Chen, T Yu - 2024 2nd International Conference On …, 2024 - ieeexplore.ieee.org
Over the past several years, deep neural networks have permeated many fields of science
research and have become an essential part of real-world applications. However, when the …

Multiclass Alignment of Confidences and Softened Target Occurrences for Train-time Calibration

V Kugathasan, H Zhou, Z Izzo, G Kuruppu, S Yoon… - openreview.net
In spite of delivering remarkable predictive accuracy across many domains, including
computer vision and medical imaging, Deep Neural Networks (DNNs) are susceptible to …

Calibration Bottleneck: What Makes Neural Networks less Calibratable?

DB Wang, ML Zhang - openreview.net
While modern deep neural networks have achieved remarkable success, they have
exhibited a notable deficiency in reliably estimating uncertainty. Many existing studies …