Generalizability vs. Robustness: Investigating Medical Imaging Networks Using Adversarial Examples

Paschali, Magdalini; Conjeti, Sailesh; Navarro, Fernando A.; Navab, Nassir

doi:10.1007/978-3-030-00928-1_56

Cited by 101 publications

(80 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For NLP or sentiment analysis, saliency map can also take the form of "heat" scores over words in texts, as demonstrated by Arras et al [62] using LRP and by Karpathy et al [63]. In the medical field (see later section), Irvin et al [6], Zhao et al [44], Paschali et al [64], Couture et al [65], Li et al [66], Qin et al [67], Tang et al [68], Papanastasopoulos et al [69], and Lee et al [70] have studied methods employing saliency and visual explanations. It is noted that we also subcategorize LIME as a method that uses optimization and sensitivity as its underlying mechanisms, and many researches on interpretability span more than one subcategories.…”

Section: A Perceptive Interpretabilitymentioning

confidence: 99%

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI

Tjoa

Guan

2021

IEEE Trans. Neural Netw. Learning Syst.

1,213

547

View full text Add to dashboard Cite

Recently, artificial intelligence and machine learning in general have demonstrated remarkable performances in many tasks, from image processing to natural language processing, especially with the advent of deep learning (DL). Along with research progress, they have encroached upon many different fields and disciplines. Some of them require high level of accountability and thus transparency, for example, the medical sector. Explanations for machine decisions and predictions are thus needed to justify their reliability. This requires greater interpretability, which often means we need to understand the mechanism underlying the algorithms. Unfortunately, the blackbox nature of the DL is still unresolved, and many machine decisions are still poorly understood. We provide a review on interpretabilities suggested by different research works and categorize them. The different categories show different dimensions in interpretability research, from approaches that provide "obviously" interpretable information to the studies of complex patterns. By applying the same categorization to interpretability in medical research, it is hoped that: 1) clinicians and practitioners can subsequently approach these methods with caution; 2) insight into interpretability will be born with more considerations for medical practices; and 3) initiatives to push forward data-based, mathematically grounded, and technically grounded medical education are encouraged.

show abstract

Section: A Perceptive Interpretabilitymentioning

confidence: 99%

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI

Tjoa

Guan

2021

IEEE Trans. Neural Netw. Learning Syst.

1,213

547

View full text Add to dashboard Cite

show abstract

“…Due to the lack of annotated data, GANs are used to generate crafted images, for example to synthesize multiple realistic-looking retinal images from an unseen tubular structured annotation that contains the binary vessel morphology in [301] or to generate synthetic liver lesion images in [302]. In [303], adversarial examples are explored in medical imaging and leveraged in a constructive fashion to benchmark model performance not only on clean and noisy but also on adversarially crafted data. In [304], a conditional GAN is explored to augment artificially generated lung nodules to improve the robustness of the progressive holistically nested network (P-HNN) model for pathological lung segmentation of CT scans.…”

Section: Emergent Architectures: the Generative Adversarial Networkmentioning

confidence: 99%

Deep Learning in the Biomedical Applications: Recent and Future Status

2019

View full text Add to dashboard Cite

Deep neural networks represent, nowadays, the most effective machine learning technology in biomedical domain. In this domain, the different areas of interest concern the Omics (study of the genome—genomics—and proteins—transcriptomics, proteomics, and metabolomics), bioimaging (study of biological cell and tissue), medical imaging (study of the human organs by creating visual representations), BBMI (study of the brain and body machine interface) and public and medical health management (PmHM). This paper reviews the major deep learning concepts pertinent to such biomedical applications. Concise overviews are provided for the Omics and the BBMI. We end our analysis with a critical discussion, interpretation and relevant open challenges.

show abstract

“…By contrast, in the embedding space created by baseline classification, a sample is more mixed with neighbours of different classes. A dispersed embedding could be one reason that classification networks are vulnerable to attacks of adversarial samples [6]. Fig.…”

Section: Evaluation Of Image Retrievalmentioning

confidence: 99%

“…This issue becomes critical in computational pathology, as pathologists need to understand the rationale of a network's decision for a certain input, if they would like to use it for diagnostic purpose. Moreover, recent studies have found deep neural networks are particularly vulnerable to adversarial examples [6]: with a small amount of permutations that are imperceptible to human, adversarial inputs can easily fool deep neural network and result in completely wrong classification, which suggest the danger of using deep neural networks without expert control.…”

Section: Introductionmentioning

confidence: 99%

Multi-task Learning of a Deep K-Nearest Neighbour Network for Histopathological Image Classification and Retrieval

Peng

Boxberg

Weichert

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Deep neural networks have achieved tremendous success in image recognition, classification and object detection. However, deep learning is often criticised for its lack of transparency and general inability to rationalize its predictions. The issue of poor model interpretability becomes critical in medical applications, as a model that is not understood and trusted by physicians is unlikely to be used in daily clinical practice. In this work, we develop a novel multi-task deep learning framework for simultaneous histopathology image classification and retrieval, leveraging on the classic concept of k-nearest neighbors to improve model interpretability. For a test image, we retrieve the most similar images from our training databases. These retrieved nearest neighbours can be used to classify the test image with a confidence score, and provide a human-interpretable explanation of our classification. Our original framework can be built on top of any existing classification network (and therefore benefit from pretrained models), by (i) adding a triplet loss function with a novel triplet sampling strategy to compare distances between samples and (ii) a Cauchy hashing loss function to accelerate neighbour searching. We evaluate our method on colorectal cancer histology slides, and show that the confidence estimates are strongly correlated with model performance. The explanations provided by nearest neighbors are intuitive and useful for expert evaluation by giving insights into understanding possible model failures, and can support clinical decision making by comparing archived images and patient records with the actual case.

show abstract

Generalizability vs. Robustness: Investigating Medical Imaging Networks Using Adversarial Examples

Cited by 101 publications

References 11 publications

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI

Deep Learning in the Biomedical Applications: Recent and Future Status

Multi-task Learning of a Deep K-Nearest Neighbour Network for Histopathological Image Classification and Retrieval

Contact Info

Product

Resources

About