From hype to hope to hard work: developing responsible AI for radiology

Rockall, Andrea

doi:10.1016/j.crad.2019.09.123

Cited by 12 publications

(6 citation statements)

References 8 publications

(8 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In medical applications such as rehabilitation, the requirement for ART AI is heightened. High-quality data as well as the responsibilities of curation, provenance and bias avoidance are critical [23]. As mentioned previously explainability and interpretability are important for engagement but also help avoid unintended consequences.…”

Section: Literature Review Of Art Ai and Its Applicability To Home-based Rehabilitation Systemsmentioning

confidence: 99%

“…As mentioned previously explainability and interpretability are important for engagement but also help avoid unintended consequences. In the medical domain, unintended consequences of AI can be significantly damaging or irreversible [23]. The benefit to patient care and society, however, can be tremendous.…”

Section: Literature Review Of Art Ai and Its Applicability To Home-based Rehabilitation Systemsmentioning

confidence: 99%

“…Current EU regulation allows for individuals to enquire about AI decisions [22]. However, regulation is not well defined for the design and development of such models, particularly in the medical applications domain [23]. Moreover, accountability and transparency are strongly interlinked with interpretability [17].…”

mentioning

confidence: 99%

See 2 more Smart Citations

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Vourganas

Stanković

2020

Sensors

View full text Add to dashboard Cite

Socioeconomic reasons post-COVID-19 demand unsupervised home-based rehabilitation and, specifically, artificial ambient intelligence with individualisation to support engagement and motivation. Artificial intelligence must also comply with accountability, responsibility, and transparency (ART) requirements for wider acceptability. This paper presents such a patient-centric individualised home-based rehabilitation support system. To this end, the Timed Up and Go (TUG) and Five Time Sit To Stand (FTSTS) tests evaluate daily living activity performance in the presence or development of comorbidities. We present a method for generating synthetic datasets complementing experimental observations and mitigating bias. We present an incremental hybrid machine learning algorithm combining ensemble learning and hybrid stacking using extreme gradient boosted decision trees and k-nearest neighbours to meet individualisation, interpretability, and ART design requirements while maintaining low computation footprint. The model reaches up to 100% accuracy for both FTSTS and TUG in predicting associated patient medical condition, and 100% or 83.13%, respectively, in predicting area of difficulty in the segments of the test. Our results show an improvement of 5% and 15% for FTSTS and TUG tests, respectively, over previous approaches that use intrusive means of monitoring such as cameras.

show abstract

Section: Literature Review Of Art Ai and Its Applicability To Home-based Rehabilitation Systemsmentioning

confidence: 99%

Section: Literature Review Of Art Ai and Its Applicability To Home-based Rehabilitation Systemsmentioning

confidence: 99%

See 1 more Smart Citation

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Vourganas

Stanković

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…A significant number of recent studies have countered this AI hype in fields such as translational medicine (Toh et al, 2019), multidisciplinary medical teams (Di Ieva, 2019), radiology (Rockall, 2020), COVID-19 (Abdulkareem and Petersen, 2021), machine vision (Marquardt, 2020), management (Holmström and Hällgren, 2021), and interaction design (Liikkanen, 2019). The advances of the last decade have indeed been significant, but AI is only capable of performing well on narrow tasks for which they can be trained over an extended period with a significant amount of relevant data (and significant number of humans working on labeling this data).…”

Section: Ai Is Not That Special (Socio-technical Critique)mentioning

confidence: 99%

Politics by Automatic Means? A Critique of Artificial Intelligence Ethics at Work

Cole

Cant²,

Spilda³

et al. 2022

Front. Artif. Intell.

View full text Add to dashboard Cite

Calls for “ethical Artificial Intelligence” are legion, with a recent proliferation of government and industry guidelines attempting to establish ethical rules and boundaries for this new technology. With few exceptions, they interpret Artificial Intelligence (AI) ethics narrowly in a liberal political framework of privacy concerns, transparency, governance and non-discrimination. One of the main hurdles to establishing “ethical AI” remains how to operationalize high-level principles such that they translate to technology design, development and use in the labor process. This is because organizations can end up interpreting ethics in an ad-hoc way with no oversight, treating ethics as simply another technological problem with technological solutions, and regulations have been largely detached from the issues AI presents for workers. There is a distinct lack of supra-national standards for fair, decent, or just AI in contexts where people depend on and work in tandem with it. Topics such as discrimination and bias in job allocation, surveillance and control in the labor process, and quantification of work have received significant attention, yet questions around AI and job quality and working conditions have not. This has left workers exposed to potential risks and harms of AI. In this paper, we provide a critique of relevant academic literature and policies related to AI ethics. We then identify a set of principles that could facilitate fairer working conditions with AI. As part of a broader research initiative with the Global Partnership on Artificial Intelligence, we propose a set of accountability mechanisms to ensure AI systems foster fairer working conditions. Such processes are aimed at reshaping the social impact of technology from the point of inception to set a research agenda for the future. As such, the key contribution of the paper is how to bridge from abstract ethical principles to operationalizable processes in the vast field of AI and new technology at work.

show abstract

“…As an increasing amount of innovative AI research has been published, a popular saying, "garbage in, garbage out," has emerged alerting researchers to pay attention to the quality of data and the reliability of methodology (Rockall, 2020). Recent studies that have attempted to apply CNN algorithms to caries detection have faced limitations.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for Caries Detection using Optical Coherence Tomography

Huang¹,

2021

Preprint

View full text Add to dashboard Cite

Early detection of dental caries has been one of the most predominant topics studied over the last few decades. Conventional examination through visual-tactile inspection and radiography can be inaccurate and destructive to teeth structure. The development of Optical Coherence Tomography (OCT) has given dentistry an alternative diagnostic technique, which has been proven by numerous studies, that it has better sensitivity, specificity, and non-invasive characteristics. The growing popularity of Artificial Intelligence (AI) also contributes to a more efficient and effective way of image-based detection and decision-making. Previous studies, which have attempted to employ AI for caries assessment, did not incorporate high-quality data. Hence, they were unable to produce valid and reliable results. This study highlights the importance of high-quality data and aims to bypass this issue, by implementing an improved methodology to the automated detection and diagnosis of dental caries depending on AI. A two-phase study was carried out to explore different methods for caries detection. Initially OCT was verified, by surveying experienced clinicians, to be a better imaging technique compared to radiography. Then, our study showed that Convolutional Neural Networks (CNNs) in the scope of AI surpassed the accuracy of human clinicians. The data was preprocessed and labelled with the ground truth corresponding to Micro-CT with rigorous definition. Statistical analysis performed was mainly based on Cohen′s kappa coefficient. The results suggested that OCT (κ = .611, SD = .107) showed a higher accuracy than radiography (κ = .326, SD = .039) and CNNs (κ = .835, SD = .057) were rated higher than clinicians (κ = .538, SD = .131), both within a .05 significance. The best result was carried out by ResNet-152, concluding diagnostic accuracy to be 95.21% and sensitivity 98.85%. The improved methodology of this study hopes to pave the way for future studies in AI application in Dentistry.

show abstract

From hype to hope to hard work: developing responsible AI for radiology

Cited by 12 publications

References 8 publications

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Politics by Automatic Means? A Critique of Artificial Intelligence Ethics at Work

Deep Learning for Caries Detection using Optical Coherence Tomography

Contact Info

Product

Resources

About