article · Artificial Intelligence Review
Deep learning models used in medical diagnostics face significant challenges in maintaining performance when confronted with noisy or adversarial inputs. Factors affecting model reliability include model complexity, the quality of training data, and hyperparameter choices. Diagnostic systems are vulnerable to adversarial attacks designed to deceive them, as well as privacy attacks that attempt to extract sensitive patient data. To counter these threats, protective strategies are under consideration, such as adversarial training, input preprocessing, data augmentation, and uncertainty estimation. In addition, existing robustness evaluation metrics are assessed alongside software packages and extensions designed to bolster reliability within standard frameworks like TensorFlow and PyTorch. Addressing current research gaps and improving these defensive techniques remains essential for developing dependable, stable, and trustworthy artificial intelligence systems for healthcare settings.
Accurate medical diagnosis relies on artificial intelligence systems that do not fail when processing unusual or corrupted data. If diagnostic algorithms are vulnerable to manipulation or privacy breaches, patient safety and confidential data are put at risk. Identifying defensive methods and reliability tools ensures healthcare technologies remain stable, secure, and dependable when deployed in clinical environments.
This work informs developers and healthcare technology providers seeking to deploy secure artificial intelligence diagnostics. By identifying defensive techniques like adversarial training and tools within standard frameworks, it supports the creation of hardened diagnostic software. However, as an evaluation of existing literature, metrics, and software extensions, the work describes early-stage research rather than a direct commercial implementation.
AI-generated from the published abstract. Always read the original work before citing.
The current study investigates the robustness of deep learning models for accurate medical diagnosis systems with a specific focus on their ability to maintain performance in the presence of adversarial or noisy inputs. We examine factors that may influence model reliability, including model complexity, training data quality, and hyperparameters; we also examine security concerns related to adversarial attacks that aim to deceive models along with privacy attacks that seek to extract sensitive information. Researchers have discussed various defenses to these attacks to enhance model robustness, such as adversarial training and input preprocessing, along with mechanisms like data augmentation and uncertainty estimation. Tools and packages that extend the reliability features of deep learning frameworks such as TensorFlow and PyTorch are also being explored and evaluated. Existing evaluation metrics for robustness are additionally being discussed and evaluated. This paper concludes by discussing limitations in the existing literature and possible future research directions to continue enhancing the status of this research topic, particularly in the medical domain, with the aim of ensuring that AI systems are trustworthy, reliable, and stable.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1007/s10462-024-11005-9
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
