Checklist for responsible deep learning modeling of medical images based on COVID-19 detection studies

Hryniewska, Weronika; Bombiński, Przemysław; Szatkowski, Patryk; Tomaszewska, Paulina; Przelaskowski, Artur; Biecek, Przemysław

doi:10.1016/j.patcog.2021.108035

Cited by 45 publications

(42 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hryniewska et al. [65] presents a checklist for the development of an ML model for lung image analysis, pointing out the urgent need for better quality and quantity of image data. One of the topics in the checklist is data augmentation, which includes image visibility, the inclusion of areas of interest, and sensible transformations.…”

Section: Next Steps and Challengesmentioning

confidence: 99%

AI-Based human audio processing for COVID-19: A comprehensive overview

Deshpande

Batliner

Schuller

2022

Pattern Recognition

View full text Add to dashboard Cite

The Coronavirus (COVID-19) pandemic impelled several research efforts, from collecting COVID-19 patients’ data to screening them for virus detection. Some COVID-19 symptoms are related to the functioning of the respiratory system that influences speech production; this suggests research on identifying markers of COVID-19 in speech and other human generated audio signals. In this article, we give an overview of research on human audio signals using ’Artificial Intelligence’ techniques to screen, diagnose, monitor, and spread the awareness about COVID-19. This overview will be useful for developing automated systems that can help in the context of COVID-19, using non-obtrusive and easy to use bio-signals conveyed in human non-speech and speech audio productions.

show abstract

Section: Next Steps and Challengesmentioning

confidence: 99%

AI-Based human audio processing for COVID-19: A comprehensive overview

Deshpande

Batliner

Schuller

2022

Pattern Recognition

View full text Add to dashboard Cite

show abstract

“…For example, machine learning systems are developed to diagnose patients based on various measurements of gene expression, even when the exact molecular processes involved in the disease are only partially known or understood. In such situations, the machine learning literature is content to measure the "quality" of a diagnostic system by measuring the accuracy of its predictions on a limited data set that was not used during the development (training, learning, adaptation, and tuning) of the system, i.e., the so-called "test data" [26,27]. That is, the algorithm is trained on a carefully selected training and test data set to develop the ability to perform a specific task, such as making a clinical diagnosis.…”

Section: Skill-based Systemsmentioning

confidence: 99%

Explainable Artificial Intelligence (XAI) in Biomedicine: Making AI Decisions Trustworthy for Physicians and Patients

2021

View full text Add to dashboard Cite

The use of artificial intelligence (AI) systems in biomedical and clinical settings can disrupt the traditional doctor–patient relationship, which is based on trust and transparency in medical advice and therapeutic decisions. When the diagnosis or selection of a therapy is no longer made solely by the physician, but to a significant extent by a machine using algorithms, decisions become nontransparent. Skill learning is the most common application of machine learning algorithms in clinical decision making. These are a class of very general algorithms (artificial neural networks, classifiers, etc.), which are tuned based on examples to optimize the classification of new, unseen cases. It is pointless to ask for an explanation for a decision. A detailed understanding of the mathematical details of an AI algorithm may be possible for experts in statistics or computer science. However, when it comes to the fate of human beings, this “developer’s explanation” is not sufficient. The concept of explainable AI (XAI) as a solution to this problem is attracting increasing scientific and regulatory interest. This review focuses on the requirement that XAIs must be able to explain in detail the decisions made by the AI to the experts in the field.

show abstract

“…Checking the responsibility of a network to classify lung lesions could be an example of using the mask loading functionality. The neural network was trained on the data with lesions label, and then validated on an external database, as recommended by Hryniewska et al [2021]. For external validation, the dataset for lesions detection was chosen, so the database contained not only the names of lesions present on the images but also their location.…”

Section: Different Types Of Superpixels' Selectionmentioning

confidence: 99%

LIMEcraft: Handcrafted superpixel selection and inspection for Visual eXplanations

Hryniewska¹,

Adrianna²,

Biecek³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

The increased interest in deep learning applications, and their hard-to-detect biases result in the need to validate and explain complex models. However, current explanation methods are limited as far as both the explanation of the reasoning process and prediction results are concerned. They usually only show the location in the image that was important for model prediction. The lack of possibility to interact with explanations makes it difficult to verify and understand exactly how the model works. This creates a significant risk when using the model. It is compounded by the fact that explanations do not take into account the semantic meaning of the explained objects. To escape from the trap of static explanations, we propose an approach called LIMEcraft that allows a user to interactively select semantically consistent areas and thoroughly examine the prediction for the image instance in case of many image features. Experiments on several models showed that our method improves model safety by inspecting model fairness for image pieces that may indicate model bias.

show abstract

Checklist for responsible deep learning modeling of medical images based on COVID-19 detection studies

Cited by 45 publications

References 51 publications

AI-Based human audio processing for COVID-19: A comprehensive overview

AI-Based human audio processing for COVID-19: A comprehensive overview

Explainable Artificial Intelligence (XAI) in Biomedicine: Making AI Decisions Trustworthy for Physicians and Patients

LIMEcraft: Handcrafted superpixel selection and inspection for Visual eXplanations

Contact Info

Product

Resources

About