Resilience and Resilient Systems of Artificial Intelligence: Taxonomy, Models and Methods

Москаленко, В`ячеслав Васильович; Kharchenko, Vyacheslav; Moskalenko, Alona; Kuzikov, Borys Olehovych

doi:10.3390/a16030165

Cited by 11 publications

(6 citation statements)

References 115 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Solving problems related to ensuring AI resilience, i.e., security, trustworthiness, robustness and rapid adaptation to changes, relies on specific expertise not related to a particular application area ( 13 ). The main difficulty in separating the tasks is the lack of universality of methods for ensuring resilience and the lack of a complete understanding of the compatibility of methods that provide different aspects of resilience and protection against different types of disturbances ( 16 ). Determining the compatibility of methods and combining them can increase flexibility and resilience depending on requirements and constraints.…”

Section: Discussionmentioning

confidence: 99%

“…The resilience of AI systems should, on the one hand, take into account traditional models and principles ( 47 , 48 ) of its provision, which are based on evolutionary mechanisms for taking into account and adapting to changes in requirements, parameters of the information and physical environments, tolerance to uncertain failures, taking into account cyberattacks ( 49 , 50 ). On the other hand, as it was mentioned in ( 16 ) the actual models and means of artificial intelligence have the potential of natural resilience, which should be used and which is used. This study is the next step in improving both of these approaches.…”

Section: Discussionmentioning

confidence: 99%

“…Disturbances in an intelligent system can be caused by both intentional attacks and natural causes. Examples of natural disturbances include noise in the data, sudden hardware faults (memory faults), and data drift over time ( 16 ). Intentional attacks can also include fault injections and data manipulation in the form of so-called adversarial attacks.…”

Section: Architecting Resilient Mlops-based Medical Diagnostic Systemmentioning

confidence: 99%

See 2 more Smart Citations

Resilience-aware MLOps for AI-based medical diagnostic system

Moskalenko,

Kharchenko

2024

Front. Public Health

Self Cite

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: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

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

Self Cite

View full text Add to dashboard Cite

show abstract

“… is a set of parameters of meta-tuners; INTRODUCTION AIS are vulnerable to various types of disturbances. The most studied types of disturbances are fault injection, adversarial attacks, drift, and out-of-distribution data [1]. These disturbances can lead to financial and human life losses in safety-critical applications.…”

Section: Nomenclaturementioning

confidence: 99%

Model-Agnostic Meta-Learning for Resilience Optimization of Artificial Intelligence System

Москаленко

2023

RIC

View full text Add to dashboard Cite

Context. The problem of optimizing the resilience of artificial intelligence systems to destructive disturbances has not yet been fully solved and is quite relevant for safety-critical applications. The task of optimizing the resilience of an artificial intelligence system to disturbing influences is a high-level task in relation to efficiency optimization, which determines the prospects of using the ideas and methods of meta-learning to solve it. The object of current research is the process of meta-learning aimed at optimizing the resilience of an artificial intelligence system to destructive disturbances. The subjects of the study are architectural add-ons and the meta-learning method which optimize resilience to adversarial attacks, fault injection, and task changes. Objective. Stated research goal is to develop an effective meta-learning method for optimizing the resilience of an artificial intelligence system to destructive disturbances. Method. The resilience optimization is implemented by combining the ideas and methods of adversarial learning, fault-tolerant learning, model-agnostic meta-learning, few-shot learning, gradient optimization methods, and probabilistic gradient approximation strategies. The choice of architectural add-ons is based on parameter-efficient knowledge transfer designed to save resources and avoid the problem of catastrophic forgetting. Results. A model-agnostic meta-learning method for optimizing the resilience of artificial intelligence systems based on gradient meta-updates or meta-updates using an evolutionary strategy has been developed. This method involves the use of tuner and metatuner blocks that perform parallel correction of the building blocks of a original deep neural network. The ability of the proposed approach to increase the efficiency of perturbation absorption and increase the integral resilience indicator of the artificial intelligence system is experimentally tested on the example of the image classification task. The experiments were conducted on a model with the ResNet-18 architecture, with an add-on in the form of tuners and meta-tuners with the Conv-Adapter architecture. In this case, CIFAR-10 is used as a base set on which the model was trained, and CIFAR-100 is used as a set for generating samples on which adaptation is performed using a few-shot learning scenarios. We compare the resilience of the artificial intelligence system after pre-training tuners and meta-tuners using the adversarial learning algorithm, the fault-tolerant learning algorithm, the conventional model-agnostic meta-learning algorithm, and the proposed meta-learning method for optimizing resilience. Also, the meta-learning algorithms with meta-gradient updating and meta-updating based on the evolutionary strategy are compared on the basis of the integral resilience indicator. Conclusions. It has been experimentally confirmed that the proposed method provides a better resilience to random bit-flip injection compared to fault injection training by an average of 5%. Also, the proposed method provides a better resilience to Ladversarial evasion attacks compared to adversarial training by an average of 4.8%. In addition, an average 4.8% increase in the resilience to task changes is demonstrated compared to conventional fine-tuning of tuners. Moreover, meta-learning with an evolutionary strategy provides, on average, higher values of the resilience indicator. On the downside, this meta-learning method requires more iterations.

show abstract

“…Many researchers and designers conduct research to improve the technical characteristics of GRS, such as weight, energy consumption, speed, reliability, and accuracy of gun firing. Scientific research into this area is important because the advantage over the enemy will be delivered by those GRSs that are endowed with advanced «artificial intelligence» [9] and improved operational properties. The results of military GRS research can be used to improve civilian robots, the use of which is rapidly increasing every year [5].…”

Section: Introductionmentioning

confidence: 99%

Design and research of the ground robotic system structure for weapons remote control

Dehtiarov,

Leontiev,

Miroshnychenko

et al. 2023

EEJET

View full text Add to dashboard Cite

During hostilities, ground robotic systems play an important role in minimizing losses of servicemen and suspending the combat capabilities of troops. For firing, robotic complexes are equipped with gun turrets. Researchers are conducting research to improve the performance, reliability and firing accuracy of such turrets. This work describes the design and research of an experimental sample of a ground robotic system, which is equipped with a turret for controlling the position of a machine gun. The description and results of experimental studies of dynamic loads during robot movement at different speeds and road conditions are presented. It was established that the values of the maximum accelerations that must be worked out by the stabilization system during operation for the experimental design of the robot do not exceed 20 rad/s2. The possibility of using counterweights was considered to reduce the torque of the turret guidance drive while reducing the dimensions of the robotic system structure. The description of the experimental module equipped with a control and measurement system and the results of experimental studies on determining the power of the turret drives during the manipulation of the structure are presented. A procedure of dynamic analysis and the results of modeling the movement of the gun turret in the ANSYS software package are presented. The proposed method for designing the structure ensures the determination of the impact on the structure of the complex shape of loads caused by its manipulation, to compensate for the exciting loads when the robotic system is moved over the terrain. With the help of this method, it is possible to determine and minimize the power, and therefore the energy consumption, of azimuth and lifting electric drives at the design stage

show abstract

Resilience and Resilient Systems of Artificial Intelligence: Taxonomy, Models and Methods

Cited by 11 publications

References 115 publications

Resilience-aware MLOps for AI-based medical diagnostic system

Resilience-aware MLOps for AI-based medical diagnostic system

Model-Agnostic Meta-Learning for Resilience Optimization of Artificial Intelligence System

Design and research of the ground robotic system structure for weapons remote control

Contact Info

Product

Resources

About