Building Trust in Deep Learning System towards Automated Disease Detection

Though deep learning systems have achieved high accuracy in detecting diseases from medical images, few such systems have been deployed in highly automated disease screening settings due to lack of trust in how well these systems can generalize to out-of-datasets. We propose to use uncertainty estimates of the deep learning system’s prediction to know when to accept or to disregard its prediction. We evaluate the effectiveness of using such estimates in a real-life application for the screening of diabetic retinopathy. We also generate visual explanation of the deep learning system to convey the pixels in the image that influences its decision. Together, these reveal the deep learning system’s competency and limits to the human, and in turn the human can know when to trust the deep learning system.

Download Full-text

DIABETIC RETINOPATHY, CLASSIFIED USING THE LESION-AWARE DEEP LEARNING SYSTEM, PREDICTS DIABETIC END-STAGE RENAL DISEASE IN CHINESE PATIENTS

Endocrine Practice ◽

10.4158/ep-2019-0512 ◽

2020 ◽

Vol 26 (4) ◽

pp. 429-443 ◽

Cited By ~ 3

Author(s):

Lijun Zhao ◽

Honghong Ren ◽

Junlin Zhang ◽

Yana Cao ◽

Yiting Wang ◽

...

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

Renal Disease ◽

End Stage Renal Disease ◽

Learning System ◽

Chinese Patients ◽

Glomerular Lesions ◽

Stage Renal Disease ◽

End Stage

Objective: To characterize the relationship between diabetic retinopathy (DR) and diabetic nephropathy (DN) in Chinese patients and to determine whether the severity of DR predicts end-stage renal disease (ESRD). Methods: Bilateral fundic photographs of 91 Chinese type 2 diabetic patients with biopsy-confirmed DN, not in ESRD stage, were obtained at the time of renal biopsy in this longitudinal study. The baseline severity of DR was determined using the Lesion-aware Deep Learning System (RetinalNET) in an open framework for deep learning and was graded using the Early Treatment Diabetic Retinopathy Study severity scale. Cox proportional hazard models were used to estimate the hazard ratio (HR) for the effect of the severity of diabetic retinopathy on ESRD. Results: During a median follow-up of 15 months, 25 patients progressed to ESRD. The severity of retinopathy at the time of biopsy was a prognostic factor for progression to ESRD (HR 2.18, 95% confidence interval 1.05 to 4.53, P = .04). At baseline, more severe retinopathy was associated with poor renal function, and more severe glomerular lesions. However, 30% of patients with mild retinopathy and severe glomerular lesions had higher low-density lipo-protein-cholesterol and more severe proteinuria than those with mild glomerular lesions. Additionally, 3% of patients with severe retinopathy and mild glomerular changes were more likely to have had diabetes a long time than those with severe glomerular lesions. Conclusion: Although the severity of DR predicted diabetic ESRD in patients with type 2 diabetes mellitus and DN, the severities of DR and DN were not always consistent, especially in patients with mild retinopathy or microalbuminuria. Abbreviations: CI = confidence interval; DM = diabetic mellitus; DN = diabetic nephropathy; DR = diabetic retinopathy; eGFR = estimated glomerular filtration rate; ESRD = end-stage renal disease; HbA1c = hemoglobin A1c; HR = hazard ratio; NPDR = nonproliferative diabetic retinopathy; PDR = proliferative diabetic retinopathy; SBP = systolic blood pressure; T2DM = type 2 diabetes mellitus; VEGF = vascular endothelial growth factor

Download Full-text

OCTAI: Smartphone-based Optical Coherence Tomography Image Analysis System

10.31224/osf.io/gq89c ◽

2021 ◽

Author(s):

Adrit Rao ◽

Harvey A. Fishman

Keyword(s):

Optical Coherence Tomography ◽

Image Analysis ◽

Deep Learning ◽

Early Stage ◽

Real Life ◽

Learning System ◽

Optical Coherence Tomography Image ◽

Image Analysis System ◽

Optical Coherence ◽

Oct Image

Identifying diseases in Optical Coherence Tomography (OCT) images using Deep Learning models and methods is emerging as a powerful technique to enhance clinical diagnosis. Identifying macular diseases in the eye at an early stage and preventing misdiagnosis is crucial. The current methods developed for OCT image analysis have not yet been integrated into an accessible form-factor that can be utilized in a real-life scenario by Ophthalmologists. Additionally, current methods do not employ robust multiple metric feedback. This paper proposes a highly accurate smartphone-based Deep Learning system, OCTAI, that allows a user to take an OCT picture and receive real-time feedback through on-device inference. OCTAI analyzes the input OCT image in three different ways: (1) full image analysis, (2) quadrant based analysis, and (3) disease detection based analysis. With these three analysis methods, along with an Ophthalmologist's interpretation, a robust diagnosis can potentially be made. The ultimate goal of OCTAI is to assist Ophthalmologists in making a diagnosis through a digital second opinion and enabling them to cross-check their diagnosis before making a decision based on purely manual analysis of OCT images. OCTAI has the potential to allow Ophthalmologists to improve their diagnosis and may reduce misdiagnosis rates, leading to faster treatment of diseases.

Download Full-text

Machine learning for human learners: opportunities, issues, tensions and threats

Educational Technology Research and Development ◽

10.1007/s11423-020-09858-2 ◽

2020 ◽

Author(s):

Mary E. Webb ◽

Andrew Fluck ◽

Johannes Magenheim ◽

Joyce Malyn-Smith ◽

Juliet Waters ◽

...

Keyword(s):

Machine Learning ◽

Deep Learning ◽

Ethical Issues ◽

Practical Experience ◽

Learning Systems ◽

Learning System ◽

Adaptive Behaviour ◽

Recent Developments ◽

Key Aspects ◽

School Curricula

AbstractMachine learning systems are infiltrating our lives and are beginning to become important in our education systems. This article, developed from a synthesis and analysis of previous research, examines the implications of recent developments in machine learning for human learners and learning. In this article we first compare deep learning in computers and humans to examine their similarities and differences. Deep learning is identified as a sub-set of machine learning, which is itself a component of artificial intelligence. Deep learning often depends on backwards propagation in weighted neural networks, so is non-deterministic—the system adapts and changes through practical experience or training. This adaptive behaviour predicates the need for explainability and accountability in such systems. Accountability is the reverse of explainability. Explainability flows through the system from inputs to output (decision) whereas accountability flows backwards, from a decision to the person taking responsibility for it. Both explainability and accountability should be incorporated in machine learning system design from the outset to meet social, ethical and legislative requirements. For students to be able to understand the nature of the systems that may be supporting their own learning as well as to act as responsible citizens in contemplating the ethical issues that machine learning raises, they need to understand key aspects of machine learning systems and have opportunities to adapt and create such systems. Therefore, some changes are needed to school curricula. The article concludes with recommendations about machine learning for teachers, students, policymakers, developers and researchers.

Download Full-text

Solving Current Limitations of Deep Learning Based Approaches for Plant Disease Detection

Symmetry ◽

10.3390/sym11070939 ◽

2019 ◽

Vol 11 (7) ◽

pp. 939 ◽

Cited By ~ 28

Author(s):

Marko Arsenovic ◽

Mirjana Karanovic ◽

Srdjan Sladojevic ◽

Andras Anderla ◽

Darko Stefanovic

Keyword(s):

Deep Learning ◽

Plant Disease ◽

Real Life ◽

Weather Conditions ◽

Plant Diseases ◽

Disease Classification ◽

Generative Adversarial Networks ◽

Disease Detection ◽

Single Leaf ◽

The Impact

Plant diseases cause great damage in agriculture, resulting in significant yield losses. The recent expansion of deep learning methods has found its application in plant disease detection, offering a robust tool with highly accurate results. The current limitations and shortcomings of existing plant disease detection models are presented and discussed in this paper. Furthermore, a new dataset containing 79,265 images was introduced with the aim to become the largest dataset containing leaf images. Images were taken in various weather conditions, at different angles, and daylight hours with an inconsistent background mimicking practical situations. Two approaches were used to augment the number of images in the dataset: traditional augmentation methods and state-of-the-art style generative adversarial networks. Several experiments were conducted to test the impact of training in a controlled environment and usage in real-life situations to accurately identify plant diseases in a complex background and in various conditions including the detection of multiple diseases in a single leaf. Finally, a novel two-stage architecture of a neural network was proposed for plant disease classification focused on a real environment. The trained model achieved an accuracy of 93.67%.

Download Full-text

An Effective Method for Detecting and Classifying Diabetic Retinopathy Lesions Based on Deep Learning

Computational and Mathematical Methods in Medicine ◽

10.1155/2021/9928899 ◽

2021 ◽

Vol 2021 ◽

pp. 1-13

Author(s):

Abdüssamed Erciyas ◽

Necaattin Barışçı

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

Transfer Learning ◽

Vision Loss ◽

Diabetic Patients ◽

Disease Detection ◽

Early Vision ◽

Second Stage ◽

22Nd Century ◽

Harmful Effects

Diabetic retinopathy occurs as a result of the harmful effects of diabetes on the eyes. Diabetic retinopathy is also a disease that should be diagnosed early. If not treated early, vision loss may occur. It is estimated that one third of more than half a million diabetic patients will have diabetic retinopathy by the 22nd century. Many effective methods have been proposed for disease detection with deep learning. In this study, unlike other studies, a deep learning-based method has been proposed in which diabetic retinopathy lesions are detected automatically and independently of datasets, and the detected lesions are classified. In the first stage of the proposed method, a data pool is created by collecting diabetic retinopathy data from different datasets. With Faster RCNN, lesions are detected, and the region of interests are marked. The images obtained in the second stage are classified using the transfer learning and attention mechanism. The method tested in Kaggle and MESSIDOR datasets reached 99.1% and 100% ACC and 99.9% and 100% AUC, respectively. When the obtained results are compared with other results in the literature, it is seen that more successful results are obtained.

Download Full-text

Technical and imaging factors influencing performance of deep learning systems for diabetic retinopathy

npj Digital Medicine ◽

10.1038/s41746-020-0247-1 ◽

2020 ◽

Vol 3 (1) ◽

Cited By ~ 2

Author(s):

Michelle Y. T. Yip ◽

Gilbert Lim ◽

Zhan Wei Lim ◽

Quang D. Nguyen ◽

Crystal C. Y. Chong ◽

...

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

Learning Systems ◽

Factors Influencing

Download Full-text

Evaluation of a deep learning system for the joint automated detection of diabetic retinopathy and age‐related macular degeneration

Acta Ophthalmologica ◽

10.1111/aos.14306 ◽

2019 ◽

Vol 98 (4) ◽

pp. 368-377 ◽

Cited By ~ 4

Author(s):

Cristina González‐Gonzalo ◽

Verónica Sánchez‐Gutiérrez ◽

Paula Hernández‐Martínez ◽

Inés Contreras ◽

Yara T. Lechanteur ◽

...

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

Macular Degeneration ◽

Automated Detection ◽

Learning System ◽

Age Related Macular Degeneration ◽

Age Related

Download Full-text

Automatic Multi-task Learning System for Abnormal Network Traffic Detection

International Journal of Emerging Technologies in Learning (iJET) ◽

10.3991/ijet.v13i04.8466 ◽

2018 ◽

Vol 13 (04) ◽

pp. 4 ◽

Cited By ~ 12

Author(s):

He Huang ◽

Haojiang Deng ◽

Jun Chen ◽

Luchao Han ◽

Wei Wang

Keyword(s):

Deep Learning ◽

Network Traffic ◽

Learning Systems ◽

Learning System ◽

Traffic Classification ◽

Learning Methods ◽

Web 3.0 ◽

Task Learning ◽

Knowledge Models ◽

End To End

Since the last decade of the 20th century, the Internet had become flourishing, which drew great interest in the detection of abnormal network traffic. Particular-ly, it’s impossible to manually detect the abnormal patterns from enormous traffic flow in real time. Therefore, multiple machine learning methods are adopted to solve this learning problem. Those methods differ in mathematical models, knowledge models, application scenarios and target flows. In recent years, as a consequence of the technological breakthrough of Web 3.0, the traditional types of traffic classifiers are getting outdated and people start to focus on deep learning methods. Deep learning provides the potential for end-to-end learning systems to automatically learn the abnormal patterns without massive feature engineering, saving plenty of detecting time. In this study, to further save both memory and times of learning systems, we propose a novel multi-task learning system based on convolutional neural network, which can simultaneously solve the tasks of malware detection, VPN-capsulation recognition and Trojan classification. To the best of our knowledge, it’s the first time to apply an end-to-end multi-task learn-ing system in traffic classification. In order to validate this method, we establish experiments on public malware dataset CTU-13 and VPN traffic dataset ISCX. Our system found a synergy among all these tasks and managed to achieve the state-of-the-art output for most of the experiments.

Download Full-text