Uncertainty-driven ensembles of multi-scale deep architectures for image classification

Arco, Juan E.; Ortíz, Andrés; Ramı́rez, Javier; Martínez-Murcia, Francisco Jesús; Zhang, Yudong; Górriz, Juan Manuel

doi:10.1016/j.inffus.2022.08.010

Cited by 36 publications

(12 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach proves especially valuable in scenarios where the model demonstrates uncertainty in its predictions. For example, in tasks like image classification or object detection, where a single input may have multiple plausible labels, an ensemble can capture this diversity, leading to a more comprehensive understanding of the data [71][72][73].…”

Section: Dementioning

confidence: 99%

Quantifying predictive uncertainty in damage classification for nondestructive evaluation using Bayesian approximation and deep learning

Li,

Deng

2024

Inverse Problems

View full text Add to dashboard Cite

Magnetic flux leakage (MFL), a widely used Nondestructive Evaluation (NDE) method, for inspecting pipelines to prevent potential long-term failures. During field testing, uncertainties can affect the accuracy of the inspection and the decision-making process regarding damage conditions. Therefore, it is essential to identify and quantify these uncertainties to ensure the reliability of the inspection. This study focuses on the uncertainties that arise during the inverse NDE process due to the dynamic magnetization process, which is affected by the relative motion of the MFL sensor and the material being tested. Specifically, the study investigates the uncertainties caused by sensing liftoff, which can affect the output signal of the sensing system. Due to the complexity of describing the forward uncertainty propagation process, this study compared two typical machine learning-based approximate Bayesian inference methods, Convolutional Neural Network (CNN) and Deep Ensemble (DE), to address the input uncertainty from the MFL response data. Besides, an Autoencoder method is applied to tackle the lack of experimental data for the training model by augmenting the dataset, which is constructed with the pre-trained model based on transfer learning. Prior knowledge learned from large simulated MFL signals can fine-tune the Autoencoder model which enhances the subsequent learning process on experimental MFL data with faster generalization. The augmented data from the fine-tuned Autoencoder is further applied for machine learning-based defect size classification. This study conducted prediction accuracy and uncertainty analysis with calibration, which can evaluate the prediction performance and reveal the relation between the liftoff uncertainty and prediction accuracy. Further, to strengthen the trustworthiness of the prediction results, the decision-making process guided by uncertainty is applied to provide valuable insights into the reliability of the final prediction results. Overall, the proposed framework for uncertainty quantification offers valuable insights into the assessment of reliability in MFL-based decision-making and inverse problems.

show abstract

Section: Dementioning

confidence: 99%

Quantifying predictive uncertainty in damage classification for nondestructive evaluation using Bayesian approximation and deep learning

Li,

Deng

2024

Inverse Problems

View full text Add to dashboard Cite

show abstract

“…To accommodate the demands of multi-scale deep learning in the university financial system, a redesign of the system hardware is imperative. The hardware redesign aims to facilitate the fully interconnected mode required by the multilayer perceptron network in multi-scale deep learning [15][16][17][18]. Crucially, the hardware components crucial for achieving this mode include sensors and coordinators.…”

Section: Hardware Design Of University Financial Systemmentioning

confidence: 99%

Innovative Financial Management in Higher Education: A Multi-Scale Deep Learning Approach for Risk Reduction and Quality Enhancement

Hongbin Yue

2024

jes

View full text Add to dashboard Cite

This paper presents a novel university financial management system leveraging multi-scale deep learning. With rising college enrollment and teaching complexities, traditional financial models require adaptation to mitigate risks and improve management quality. The system integrates hardware and software innovations: multiple sensors enhance data scanning, coordinated by a central coordinator, ensuring comprehensive financial database coverage. Software-wise, a structured database establishes attribute-based financial connections, crucial for weight assignment. Employing a multilayer perceptual network topology, a full interconnection model based on multi-scale deep learning facilitates profound data extraction. Experimental evaluations demonstrate the system's superior financial risk assessment capabilities compared to traditional approaches, extracting a broader spectrum of financial parameters for comprehensive risk warnings. By embracing multi-scale deep learning, this system promises significant advancements in university financial management, enhancing adaptability and risk mitigation in college finance departments.

show abstract

“…DL models have made significant advances in a variety of fields including, but not limited to, deep fakes [ 22 , 23 ], satellite image analysis [ 24 ], image classification [ 25 , 26 ], the optimization of artificial neural networks [ 27 , 28 ], the processing of natural language [ 29 , 30 ], fin-tech [ 31 ], intrusion detection [ 32 ], steganography [ 33 ], and biomedical image analysis [ 14 , 34 ]. CNNs have recently surfaced as one of the most commonly used techniques for plant disease identification [ 35 , 36 ].…”

Section: Related Workmentioning

confidence: 99%

BotanicX-AI: Identification of Tomato Leaf Diseases Using an Explanation-Driven Deep-Learning Model

et al. 2023

View full text Add to dashboard Cite

Early and accurate tomato disease detection using easily available leaf photos is essential for farmers and stakeholders as it help reduce yield loss due to possible disease epidemics. This paper aims to visually identify nine different infectious diseases (bacterial spot, early blight, Septoria leaf spot, late blight, leaf mold, two-spotted spider mite, mosaic virus, target spot, and yellow leaf curl virus) in tomato leaves in addition to healthy leaves. We implemented EfficientNetB5 with a tomato leaf disease (TLD) dataset without any segmentation, and the model achieved an average training accuracy of 99.84% ± 0.10%, average validation accuracy of 98.28% ± 0.20%, and average test accuracy of 99.07% ± 0.38% over 10 cross folds.The use of gradient-weighted class activation mapping (GradCAM) and local interpretable model-agnostic explanations are proposed to provide model interpretability, which is essential to predictive performance, helpful in building trust, and required for integration into agricultural practice.

show abstract

Uncertainty-driven ensembles of multi-scale deep architectures for image classification

Cited by 36 publications

References 97 publications

Quantifying predictive uncertainty in damage classification for nondestructive evaluation using Bayesian approximation and deep learning

Quantifying predictive uncertainty in damage classification for nondestructive evaluation using Bayesian approximation and deep learning

Innovative Financial Management in Higher Education: A Multi-Scale Deep Learning Approach for Risk Reduction and Quality Enhancement

BotanicX-AI: Identification of Tomato Leaf Diseases Using an Explanation-Driven Deep-Learning Model

Contact Info

Product

Resources

About