Continuous design control for machine learning in certified medical systems

Stirbu, Vlad; Granlund, Tuomas; Mikkonen, Tommi

doi:10.1007/s11219-022-09601-5

Cited by 7 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Machine learning operations is the process of automating the lifecycle of machine learning models. It involves four main stages ( Figure 1 ) ( 1 , 2 ): Data Preparation—gathering, cleaning, and transforming data for further model training. Model Development, Training and Evaluation—building the architecture, training, and testing the model on prepared data.…”

Section: Architecting Resilient Mlops-based Medical Diagnostic Systemmentioning

confidence: 99%

See 1 more Smart Citation

Resilience-aware MLOps for AI-based medical diagnostic system

Moskalenko,

Kharchenko

2024

Front. Public Health

View full text Add to dashboard Cite

BackgroundThe healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences.MethodsPost-hoc resilience optimization, post-hoc predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for post-hoc calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode.ResultsThe structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system’s life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters.СonclusionThe main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness.

show abstract

Section: Architecting Resilient Mlops-based Medical Diagnostic Systemmentioning

confidence: 99%

“…The evolution of AI in healthcare has led to various significant advancements, many of which are integrated into existing MLOps frameworks ( 1 ). A plethora of research exists, focusing on improving data quality, model training, evaluation, and deployment in the healthcare domain ( 2 , 3 ).…”

Section: Introductionmentioning

confidence: 99%

Resilience-aware MLOps for AI-based medical diagnostic system

Moskalenko,

Kharchenko

2024

Front. Public Health

View full text Add to dashboard Cite

show abstract

“…These are just two examples of a wider range of tasks that fall under the umbrella of MLOps. Stirbu et al provide a detailed approach of continuous design control for industrial medical ML systems [67].…”

Section: Mlopsmentioning

confidence: 99%

Analysis: Flawed Datasets of Monkeypox Skin Images

2023

View full text Add to dashboard Cite

The self-proclaimed first publicly available dataset of Monkeypox skin images consists of medically irrelevant images extracted from Google and photography repositories through a process denominated web-scrapping. Yet, this did not stop other researchers from employing it to build Machine Learning (ML) solutions aimed at computer-aided diagnosis of Monkeypox and other viral infections presenting skin lesions. Neither did it stop the reviewers or editors from publishing these subsequent works in peer-reviewed journals. Several of these works claimed extraordinary performance in the classification of Monkeypox, Chickenpox and Measles, employing ML and the aforementioned dataset. In this work, we analyse the initiator work that has catalysed the development of several ML solutions, and whose popularity is continuing to grow. Further, we provide a rebuttal experiment that showcases the risks of such methodologies, proving that the ML solutions do not necessarily obtain their performance from the features relevant to the diseases at issue.

show abstract

Role of Regulatory Sandboxes and MLOps for AI-Enabled Public Sector Services

Gonzalez Torres,

Sawhney

2023

Rev Socionetwork Strat

View full text Add to dashboard Cite

This paper discusses how innovations in public sector AI-based services must comply with the Artificial Intelligence Act (AI Act) regulatory frameworks while enabling experimentation and participation of diverse stakeholders throughout the Artificial Intelligence (AI) lifecycle. The paper examines the implications of the emerging regulation, AI regulatory sandboxes and Machine Learning Operations (MLOps) as tools that facilitate compliance while enabling co-learning and active participation of multiple stakeholders. We propose a framework that fosters experimentation with automation pipelines and continuous monitoring for the deployment of future public sector AI-based services in a regulatory-compliant and technically innovative manner. AI regulatory sandboxes can be beneficial as a space for contained experimentation that goes beyond regulatory considerations to specific experimentation with the implementation of ML frameworks. While the paper presents a framework based on emerging regulations, tools and practices pertaining to the responsible use of AI, this must be validated through pilot experimentation with public and private stakeholders and regulators in different areas of high-risk AI-based services.

show abstract

Continuous design control for machine learning in certified medical systems

Cited by 7 publications

References 19 publications

Resilience-aware MLOps for AI-based medical diagnostic system

Resilience-aware MLOps for AI-based medical diagnostic system

Analysis: Flawed Datasets of Monkeypox Skin Images

Role of Regulatory Sandboxes and MLOps for AI-Enabled Public Sector Services

Contact Info

Product

Resources

About