Self-adaptive Machine Learning Systems: Research Challenges and Opportunities

Casimiro, Maria Helena; Romano, Paolo; Garlan, David; Moreno, Gabriel A.; Kang, Eunsuk; Klein, Mark

doi:10.1007/978-3-031-15116-3_7

Cited by 6 publications

(4 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…4) Planner: The Planner component strategizes adaptations based on insights from the Analyzer, specifically addressing energy efficiency and confidence score through model selection under two scenarios as explained in Algorithm 1. The algorithm navigates decision-making under two primary scenarios-high energy and low confidence-leveraging a unified score metric for model selection (lines [11][12][13][14][15][16]. The score (score m[i]), calculated for each model, combines energy efficiency and confidence inversely (line 10), guiding the selection process.…”

Section: B Mape-k Loop 1) Knowledgementioning

confidence: 99%

“…The concept of self-adaptation in software, originating from IBM's autonomic computing [21], has evolved to include Machine Learning Systems (MLS). Research in this area [14], [22], [23] categorizes self-adaptation into software design approaches (SDA) and system engineering approaches (SEA), focusing on designtime solutions. Recent studies [15], [24] explore enhancing adaptability at runtime, including unsupervised learning and model switching.…”

Section: Related Workmentioning

confidence: 99%

“…Initial studies [14] explored how ML systems adapt to environmental shifts. Our further research [15] introduced the Machine Learning Model Balancer, which dynamically switches between ML models to ensure optimal QoS, yet overlooked energy efficiency-a critical aspect given ML deployments' increasing energy demands and environmental impact.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

EcoMLS: A Self-Adaptation Approach for Architecting Green ML-Enabled Systems

Tedla,

Kulkarni,

Vaidhyanathan

2024

Preprint

View full text Add to dashboard Cite

The sustainability of Machine Learning-Enabled Systems (MLS), particularly with regard to energy efficiency, is an important challenge in their development and deployment. Self-adaptation techniques, recognized for their potential in energy savings within software systems, have yet to be extensively explored in Machine Learning-Enabled Systems (MLS), where runtime uncertainties can significantly impact model performance and energy consumption. This variability, alongside the fluctuating energy demands of ML models during operation, necessitates a dynamic approach. Addressing these challenges, we introduce EcoMLS approach, which leverages the Machine Learning Model Balancer concept to enhance the sustainability of MLS through runtime ML model switching. By adapting to monitored runtime conditions, EcoMLS optimally balances energy consumption with model confidence, demonstrating a significant advancement towards sustainable, energy-efficient machine learning solutions. Through an object detection exemplar, we illustrate the application of EcoMLS, showcasing its ability to reduce energy consumption while maintaining high model accuracy throughout its use. This research underscores the feasibility of enhancing MLS sustainability through intelligent runtime adaptations, contributing a valuable perspective to the ongoing discourse on energy-efficient machine learning.

show abstract

Section: B Mape-k Loop 1) Knowledgementioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

EcoMLS: A Self-Adaptation Approach for Architecting Green ML-Enabled Systems

Tedla,

Kulkarni,

Vaidhyanathan

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Moreover, many self-adaption mechanisms have to work in coordinated, cooperative ways within and across several system layers [28]. Thereby it can easily happen that a change triggered by one adaption mechanism triggers a whole cascade of further adaptions and may even require an alteration in a subsequent adaption mechanism and its learned model [29]. For all these reasons, the learning systems used within adaption mechanisms should themselves be flexible, easily adapting learned models to changed circumstances using only little extra data.…”

Section: A Flexibility In Computer-integrated Manufacturingmentioning

confidence: 99%

The Elements of Flexibility for Task-Performing Systems

2023

View full text Add to dashboard Cite

What makes living systems flexible so that they can react quickly and adapt easily to changing environments? This question has not only engaged biologists for decades but is also of great interest to computer scientists and engineers who seek inspiration from nature to increase the flexibility of task-performing systems such as machine learning systems, robots, or manufacturing systems. In this paper, we give a broad overview of design features of living systems that are known to promote flexibility. We call these design features the ''elements of flexibility''. Moreover, to facilitate interdisciplinary, bio-inspired research that brings the elements of flexibility to man-made task-performing systems, we introduce a general formalism for system flexibility optimization. The formalism is intended to (i) provide a common language to communicate ideas about system flexibility among researchers with different backgrounds, (ii) help to understand and compare existing research on system flexibility, e.g., in transfer learning or manufacturing flexibility, and (iii) provide a basis for a general theory of system flexibility optimization.

show abstract